{"adam_lr": 1e-06, "dt_ms": 9068.022966, "global_tokens_per_sec": 1806.788543, "grad_norm": 63.5, "loss": 5.666769, "lr_stage": "warmup", "mfu": 1.037889, "stage_name": "permutation_cycle", "status": "training", "step": 1, "tokens_per_sec_per_gpu": 1806.788543, "vram_gb": 18.430966} {"adam_lr": 1.5e-06, "dt_ms": 156.311989, "global_tokens_per_sec": 104816.016497, "grad_norm": 55.75, "loss": 5.801567, "lr_stage": "warmup", "mfu": 60.21036, "stage_name": "permutation_cycle", "status": "training", "step": 2, "tokens_per_sec_per_gpu": 104816.016497, "vram_gb": 20.709735} {"adam_lr": 2e-06, "dt_ms": 158.36978, "global_tokens_per_sec": 103454.080966, "grad_norm": 63.5, "loss": 6.341677, "lr_stage": "warmup", "mfu": 59.428011, "stage_name": "permutation_cycle", "status": "training", "step": 3, "tokens_per_sec_per_gpu": 103454.080966, "vram_gb": 20.709735} {"adam_lr": 2.5e-06, "dt_ms": 158.518076, "global_tokens_per_sec": 103357.297914, "grad_norm": 64.0, "loss": 5.477771, "lr_stage": "warmup", "mfu": 59.372416, "stage_name": "permutation_cycle", "status": "training", "step": 4, "tokens_per_sec_per_gpu": 103357.297914, "vram_gb": 20.709735} {"adam_lr": 3e-06, "dt_ms": 157.405138, "global_tokens_per_sec": 104088.088906, "grad_norm": 64.0, "loss": 5.483265, "lr_stage": "warmup", "mfu": 59.79221, "stage_name": "permutation_cycle", "status": "training", "step": 5, "tokens_per_sec_per_gpu": 104088.088906, "vram_gb": 20.709735} {"adam_lr": 3.5e-06, "dt_ms": 157.72438, "global_tokens_per_sec": 103877.409116, "grad_norm": 65.5, "loss": 5.563436, "lr_stage": "warmup", "mfu": 59.671187, "stage_name": "permutation_cycle", "status": "training", "step": 6, "tokens_per_sec_per_gpu": 103877.409116, "vram_gb": 20.709735} {"adam_lr": 4e-06, "dt_ms": 168.284416, "global_tokens_per_sec": 97358.985283, "grad_norm": 69.0, "loss": 5.286873, "lr_stage": "warmup", "mfu": 55.926754, "stage_name": "permutation_cycle", "status": "training", "step": 7, "tokens_per_sec_per_gpu": 97358.985283, "vram_gb": 20.709735} {"adam_lr": 4.5e-06, "dt_ms": 163.692236, "global_tokens_per_sec": 100090.269433, "grad_norm": 64.5, "loss": 5.924864, "lr_stage": "warmup", "mfu": 57.495708, "stage_name": "permutation_cycle", "status": "training", "step": 8, "tokens_per_sec_per_gpu": 100090.269433, "vram_gb": 20.709735} {"adam_lr": 5e-06, "dt_ms": 159.347773, "global_tokens_per_sec": 102819.134104, "grad_norm": 58.0, "loss": 5.74296, "lr_stage": "warmup", "mfu": 59.063274, "stage_name": "permutation_cycle", "status": "training", "step": 9, "tokens_per_sec_per_gpu": 102819.134104, "vram_gb": 20.709735} {"adam_lr": 5.500000000000001e-06, "dt_ms": 160.270452, "global_tokens_per_sec": 102227.202485, "grad_norm": 52.0, "loss": 5.856318, "lr_stage": "warmup", "mfu": 58.723245, "stage_name": "permutation_cycle", "status": "training", "step": 10, "tokens_per_sec_per_gpu": 102227.202485, "vram_gb": 20.709735} {"adam_lr": 6e-06, "dt_ms": 159.034252, "global_tokens_per_sec": 103021.831944, "grad_norm": 61.25, "loss": 5.820127, "lr_stage": "warmup", "mfu": 59.179711, "stage_name": "permutation_cycle", "status": "training", "step": 11, "tokens_per_sec_per_gpu": 103021.831944, "vram_gb": 20.709735} {"adam_lr": 6.5000000000000004e-06, "dt_ms": 158.955574, "global_tokens_per_sec": 103072.824589, "grad_norm": 60.5, "loss": 5.949164, "lr_stage": "warmup", "mfu": 59.209003, "stage_name": "permutation_cycle", "status": "training", "step": 12, "tokens_per_sec_per_gpu": 103072.824589, "vram_gb": 20.709735} {"adam_lr": 7e-06, "dt_ms": 163.351297, "global_tokens_per_sec": 100299.172782, "grad_norm": 58.0, "loss": 5.444772, "lr_stage": "warmup", "mfu": 57.615711, "stage_name": "permutation_cycle", "status": "training", "step": 13, "tokens_per_sec_per_gpu": 100299.172782, "vram_gb": 20.709735} {"adam_lr": 7.5e-06, "dt_ms": 167.034388, "global_tokens_per_sec": 98087.586853, "grad_norm": 47.75, "loss": 6.117317, "lr_stage": "warmup", "mfu": 56.34529, "stage_name": "permutation_cycle", "status": "training", "step": 14, "tokens_per_sec_per_gpu": 98087.586853, "vram_gb": 20.709735} {"adam_lr": 8e-06, "dt_ms": 161.423445, "global_tokens_per_sec": 101497.028673, "grad_norm": 66.5, "loss": 5.740096, "lr_stage": "warmup", "mfu": 58.303805, "stage_name": "permutation_cycle", "status": "training", "step": 15, "tokens_per_sec_per_gpu": 101497.028673, "vram_gb": 20.709735} {"adam_lr": 8.5e-06, "dt_ms": 159.978628, "global_tokens_per_sec": 102413.679806, "grad_norm": 66.5, "loss": 5.771081, "lr_stage": "warmup", "mfu": 58.830365, "stage_name": "permutation_cycle", "status": "training", "step": 16, "tokens_per_sec_per_gpu": 102413.679806, "vram_gb": 20.709735} {"adam_lr": 9e-06, "dt_ms": 160.922766, "global_tokens_per_sec": 101812.816387, "grad_norm": 55.0, "loss": 5.535097, "lr_stage": "warmup", "mfu": 58.485206, "stage_name": "permutation_cycle", "status": "training", "step": 17, "tokens_per_sec_per_gpu": 101812.816387, "vram_gb": 20.709735} {"adam_lr": 9.5e-06, "dt_ms": 160.967827, "global_tokens_per_sec": 101784.315048, "grad_norm": 61.25, "loss": 5.71219, "lr_stage": "warmup", "mfu": 58.468834, "stage_name": "permutation_cycle", "status": "training", "step": 18, "tokens_per_sec_per_gpu": 101784.315048, "vram_gb": 20.709735} {"adam_lr": 1e-05, "dt_ms": 164.876938, "global_tokens_per_sec": 99371.083743, "grad_norm": 61.75, "loss": 5.642633, "lr_stage": "warmup", "mfu": 57.08258, "stage_name": "permutation_cycle", "status": "training", "step": 19, "tokens_per_sec_per_gpu": 99371.083743, "vram_gb": 20.709735} {"adam_lr": 1.0500000000000001e-05, "dt_ms": 164.291143, "global_tokens_per_sec": 99725.400038, "grad_norm": 50.0, "loss": 5.898455, "lr_stage": "warmup", "mfu": 57.286113, "stage_name": "permutation_cycle", "status": "training", "step": 20, "tokens_per_sec_per_gpu": 99725.400038, "vram_gb": 20.709735} {"adam_lr": 1.1000000000000001e-05, "dt_ms": 162.743092, "global_tokens_per_sec": 100674.012277, "grad_norm": 60.25, "loss": 6.097636, "lr_stage": "warmup", "mfu": 57.831033, "stage_name": "permutation_cycle", "status": "training", "step": 21, "tokens_per_sec_per_gpu": 100674.012277, "vram_gb": 20.709735} {"adam_lr": 1.15e-05, "dt_ms": 159.875154, "global_tokens_per_sec": 102479.963517, "grad_norm": 61.5, "loss": 5.909586, "lr_stage": "warmup", "mfu": 58.868441, "stage_name": "permutation_cycle", "status": "training", "step": 22, "tokens_per_sec_per_gpu": 102479.963517, "vram_gb": 20.709735} {"adam_lr": 1.2e-05, "dt_ms": 159.966946, "global_tokens_per_sec": 102421.159156, "grad_norm": 54.0, "loss": 5.860123, "lr_stage": "warmup", "mfu": 58.834661, "stage_name": "permutation_cycle", "status": "training", "step": 23, "tokens_per_sec_per_gpu": 102421.159156, "vram_gb": 20.709735} {"adam_lr": 1.25e-05, "dt_ms": 160.524368, "global_tokens_per_sec": 102065.500553, "grad_norm": 62.25, "loss": 5.689479, "lr_stage": "warmup", "mfu": 58.630357, "stage_name": "permutation_cycle", "status": "training", "step": 24, "tokens_per_sec_per_gpu": 102065.500553, "vram_gb": 20.709735} {"adam_lr": 1.3000000000000001e-05, "dt_ms": 161.561251, "global_tokens_per_sec": 101410.455356, "grad_norm": 67.0, "loss": 5.955509, "lr_stage": "warmup", "mfu": 58.254074, "stage_name": "permutation_cycle", "status": "training", "step": 25, "tokens_per_sec_per_gpu": 101410.455356, "vram_gb": 20.709735} {"adam_lr": 1.3500000000000001e-05, "dt_ms": 162.26387, "global_tokens_per_sec": 100971.337463, "grad_norm": 65.0, "loss": 5.716063, "lr_stage": "warmup", "mfu": 58.001828, "stage_name": "permutation_cycle", "status": "training", "step": 26, "tokens_per_sec_per_gpu": 100971.337463, "vram_gb": 20.709735} {"adam_lr": 1.4e-05, "dt_ms": 162.96792, "global_tokens_per_sec": 100535.123535, "grad_norm": 61.25, "loss": 5.778676, "lr_stage": "warmup", "mfu": 57.75125, "stage_name": "permutation_cycle", "status": "training", "step": 27, "tokens_per_sec_per_gpu": 100535.123535, "vram_gb": 20.709735} {"adam_lr": 1.4500000000000002e-05, "dt_ms": 160.367966, "global_tokens_per_sec": 102165.042305, "grad_norm": 56.5, "loss": 6.019595, "lr_stage": "warmup", "mfu": 58.687538, "stage_name": "permutation_cycle", "status": "training", "step": 28, "tokens_per_sec_per_gpu": 102165.042305, "vram_gb": 20.709735} {"adam_lr": 1.5e-05, "dt_ms": 159.908533, "global_tokens_per_sec": 102458.572302, "grad_norm": 57.5, "loss": 6.055385, "lr_stage": "warmup", "mfu": 58.856153, "stage_name": "permutation_cycle", "status": "training", "step": 29, "tokens_per_sec_per_gpu": 102458.572302, "vram_gb": 20.709735} {"adam_lr": 1.55e-05, "dt_ms": 161.021948, "global_tokens_per_sec": 101750.104366, "grad_norm": 55.75, "loss": 5.777488, "lr_stage": "warmup", "mfu": 58.449182, "stage_name": "permutation_cycle", "status": "training", "step": 30, "tokens_per_sec_per_gpu": 101750.104366, "vram_gb": 20.709735} {"adam_lr": 1.6e-05, "dt_ms": 163.7218, "global_tokens_per_sec": 100072.195731, "grad_norm": 52.5, "loss": 5.957504, "lr_stage": "warmup", "mfu": 57.485326, "stage_name": "permutation_cycle", "status": "training", "step": 31, "tokens_per_sec_per_gpu": 100072.195731, "vram_gb": 20.709735} {"adam_lr": 1.65e-05, "dt_ms": 163.347006, "global_tokens_per_sec": 100301.807893, "grad_norm": 60.5, "loss": 5.643069, "lr_stage": "warmup", "mfu": 57.617224, "stage_name": "permutation_cycle", "status": "training", "step": 32, "tokens_per_sec_per_gpu": 100301.807893, "vram_gb": 20.709735} {"adam_lr": 1.7e-05, "dt_ms": 163.053989, "global_tokens_per_sec": 100482.055418, "grad_norm": 55.5, "loss": 5.523494, "lr_stage": "warmup", "mfu": 57.720765, "stage_name": "permutation_cycle", "status": "training", "step": 33, "tokens_per_sec_per_gpu": 100482.055418, "vram_gb": 20.709735} {"adam_lr": 1.7500000000000002e-05, "dt_ms": 160.801172, "global_tokens_per_sec": 101889.804472, "grad_norm": 59.5, "loss": 5.691195, "lr_stage": "warmup", "mfu": 58.529431, "stage_name": "permutation_cycle", "status": "training", "step": 34, "tokens_per_sec_per_gpu": 101889.804472, "vram_gb": 20.709735} {"adam_lr": 1.8e-05, "dt_ms": 162.099838, "global_tokens_per_sec": 101073.512325, "grad_norm": 53.5, "loss": 5.635554, "lr_stage": "warmup", "mfu": 58.060521, "stage_name": "permutation_cycle", "status": "training", "step": 35, "tokens_per_sec_per_gpu": 101073.512325, "vram_gb": 20.709735} {"adam_lr": 1.8500000000000002e-05, "dt_ms": 161.932707, "global_tokens_per_sec": 101177.83072, "grad_norm": 56.0, "loss": 5.872845, "lr_stage": "warmup", "mfu": 58.120446, "stage_name": "permutation_cycle", "status": "training", "step": 36, "tokens_per_sec_per_gpu": 101177.83072, "vram_gb": 20.709735} {"adam_lr": 1.9e-05, "dt_ms": 162.936449, "global_tokens_per_sec": 100554.541942, "grad_norm": 63.25, "loss": 5.937515, "lr_stage": "warmup", "mfu": 57.762404, "stage_name": "permutation_cycle", "status": "training", "step": 37, "tokens_per_sec_per_gpu": 100554.541942, "vram_gb": 20.709735} {"adam_lr": 1.95e-05, "dt_ms": 164.900303, "global_tokens_per_sec": 99357.003675, "grad_norm": 57.5, "loss": 5.442768, "lr_stage": "warmup", "mfu": 57.074492, "stage_name": "permutation_cycle", "status": "training", "step": 38, "tokens_per_sec_per_gpu": 99357.003675, "vram_gb": 20.709735} {"adam_lr": 2e-05, "dt_ms": 161.602974, "global_tokens_per_sec": 101384.272831, "grad_norm": 55.75, "loss": 5.963887, "lr_stage": "warmup", "mfu": 58.239034, "stage_name": "permutation_cycle", "status": "training", "step": 39, "tokens_per_sec_per_gpu": 101384.272831, "vram_gb": 20.709735} {"adam_lr": 2.05e-05, "dt_ms": 161.100864, "global_tokens_per_sec": 101700.261262, "grad_norm": 66.0, "loss": 5.956248, "lr_stage": "warmup", "mfu": 58.42055, "stage_name": "permutation_cycle", "status": "training", "step": 40, "tokens_per_sec_per_gpu": 101700.261262, "vram_gb": 20.709735} {"adam_lr": 2.1000000000000002e-05, "dt_ms": 160.224676, "global_tokens_per_sec": 102256.40891, "grad_norm": 61.25, "loss": 5.55124, "lr_stage": "warmup", "mfu": 58.740022, "stage_name": "permutation_cycle", "status": "training", "step": 41, "tokens_per_sec_per_gpu": 102256.40891, "vram_gb": 20.709735} {"adam_lr": 2.15e-05, "dt_ms": 161.775112, "global_tokens_per_sec": 101276.394014, "grad_norm": 66.0, "loss": 5.473292, "lr_stage": "warmup", "mfu": 58.177064, "stage_name": "permutation_cycle", "status": "training", "step": 42, "tokens_per_sec_per_gpu": 101276.394014, "vram_gb": 20.709735} {"adam_lr": 2.2000000000000003e-05, "dt_ms": 161.623001, "global_tokens_per_sec": 101371.71002, "grad_norm": 63.0, "loss": 5.759334, "lr_stage": "warmup", "mfu": 58.231817, "stage_name": "permutation_cycle", "status": "training", "step": 43, "tokens_per_sec_per_gpu": 101371.71002, "vram_gb": 20.709735} {"adam_lr": 2.25e-05, "dt_ms": 163.668156, "global_tokens_per_sec": 100104.99558, "grad_norm": 64.0, "loss": 6.038895, "lr_stage": "warmup", "mfu": 57.504168, "stage_name": "permutation_cycle", "status": "training", "step": 44, "tokens_per_sec_per_gpu": 100104.99558, "vram_gb": 20.709735} {"adam_lr": 2.3e-05, "dt_ms": 162.404776, "global_tokens_per_sec": 100883.732867, "grad_norm": 56.0, "loss": 5.630035, "lr_stage": "warmup", "mfu": 57.951504, "stage_name": "permutation_cycle", "status": "training", "step": 45, "tokens_per_sec_per_gpu": 100883.732867, "vram_gb": 20.709735} {"adam_lr": 2.3500000000000002e-05, "dt_ms": 162.221432, "global_tokens_per_sec": 100997.752424, "grad_norm": 52.5, "loss": 5.707455, "lr_stage": "warmup", "mfu": 58.017002, "stage_name": "permutation_cycle", "status": "training", "step": 46, "tokens_per_sec_per_gpu": 100997.752424, "vram_gb": 20.709735} {"adam_lr": 2.4e-05, "dt_ms": 162.546873, "global_tokens_per_sec": 100795.540931, "grad_norm": 52.0, "loss": 5.668189, "lr_stage": "warmup", "mfu": 57.900844, "stage_name": "permutation_cycle", "status": "training", "step": 47, "tokens_per_sec_per_gpu": 100795.540931, "vram_gb": 20.709735} {"adam_lr": 2.45e-05, "dt_ms": 161.814213, "global_tokens_per_sec": 101251.921674, "grad_norm": 60.0, "loss": 5.662327, "lr_stage": "warmup", "mfu": 58.163006, "stage_name": "permutation_cycle", "status": "training", "step": 48, "tokens_per_sec_per_gpu": 101251.921674, "vram_gb": 20.709735} {"adam_lr": 2.5e-05, "dt_ms": 164.295673, "global_tokens_per_sec": 99722.650414, "grad_norm": 55.0, "loss": 5.481073, "lr_stage": "warmup", "mfu": 57.284534, "stage_name": "permutation_cycle", "status": "training", "step": 49, "tokens_per_sec_per_gpu": 99722.650414, "vram_gb": 20.709735} {"adam_lr": 2.5500000000000003e-05, "dt_ms": 165.218353, "global_tokens_per_sec": 99165.738404, "grad_norm": 53.25, "loss": 5.592101, "lr_stage": "warmup", "mfu": 56.964622, "stage_name": "permutation_cycle", "status": "training", "step": 50, "tokens_per_sec_per_gpu": 99165.738404, "vram_gb": 20.709735} {"adam_lr": 2.6000000000000002e-05, "dt_ms": 162.80055, "global_tokens_per_sec": 100638.480359, "grad_norm": 58.5, "loss": 5.652993, "lr_stage": "warmup", "mfu": 57.810622, "stage_name": "permutation_cycle", "status": "training", "step": 51, "tokens_per_sec_per_gpu": 100638.480359, "vram_gb": 20.709735} {"adam_lr": 2.65e-05, "dt_ms": 161.947012, "global_tokens_per_sec": 101168.893473, "grad_norm": 70.5, "loss": 5.369175, "lr_stage": "warmup", "mfu": 58.115312, "stage_name": "permutation_cycle", "status": "training", "step": 52, "tokens_per_sec_per_gpu": 101168.893473, "vram_gb": 20.709735} {"adam_lr": 2.7000000000000002e-05, "dt_ms": 160.528183, "global_tokens_per_sec": 102063.075128, "grad_norm": 55.75, "loss": 5.891769, "lr_stage": "warmup", "mfu": 58.628964, "stage_name": "permutation_cycle", "status": "training", "step": 53, "tokens_per_sec_per_gpu": 102063.075128, "vram_gb": 20.709735} {"adam_lr": 2.7500000000000004e-05, "dt_ms": 160.535097, "global_tokens_per_sec": 102058.67934, "grad_norm": 55.25, "loss": 5.548015, "lr_stage": "warmup", "mfu": 58.626439, "stage_name": "permutation_cycle", "status": "training", "step": 54, "tokens_per_sec_per_gpu": 102058.67934, "vram_gb": 20.709735} {"adam_lr": 2.8e-05, "dt_ms": 162.776232, "global_tokens_per_sec": 100653.515702, "grad_norm": 56.0, "loss": 5.749655, "lr_stage": "warmup", "mfu": 57.819259, "stage_name": "permutation_cycle", "status": "training", "step": 55, "tokens_per_sec_per_gpu": 100653.515702, "vram_gb": 20.709735} {"adam_lr": 2.85e-05, "dt_ms": 164.636374, "global_tokens_per_sec": 99516.283369, "grad_norm": 55.25, "loss": 5.890341, "lr_stage": "warmup", "mfu": 57.165989, "stage_name": "permutation_cycle", "status": "training", "step": 56, "tokens_per_sec_per_gpu": 99516.283369, "vram_gb": 20.709735} {"adam_lr": 2.9000000000000004e-05, "dt_ms": 161.348581, "global_tokens_per_sec": 101544.121842, "grad_norm": 57.5, "loss": 5.405879, "lr_stage": "warmup", "mfu": 58.330857, "stage_name": "permutation_cycle", "status": "training", "step": 57, "tokens_per_sec_per_gpu": 101544.121842, "vram_gb": 20.709735} {"adam_lr": 2.95e-05, "dt_ms": 162.623405, "global_tokens_per_sec": 100748.10544, "grad_norm": 56.0, "loss": 5.788064, "lr_stage": "warmup", "mfu": 57.873595, "stage_name": "permutation_cycle", "status": "training", "step": 58, "tokens_per_sec_per_gpu": 100748.10544, "vram_gb": 20.709735} {"adam_lr": 3e-05, "dt_ms": 162.976265, "global_tokens_per_sec": 100529.975973, "grad_norm": 62.0, "loss": 5.798549, "lr_stage": "warmup", "mfu": 57.748293, "stage_name": "permutation_cycle", "status": "training", "step": 59, "tokens_per_sec_per_gpu": 100529.975973, "vram_gb": 20.709735} {"adam_lr": 3.0500000000000003e-05, "dt_ms": 162.005901, "global_tokens_per_sec": 101132.118428, "grad_norm": 54.0, "loss": 5.53434, "lr_stage": "warmup", "mfu": 58.094187, "stage_name": "permutation_cycle", "status": "training", "step": 60, "tokens_per_sec_per_gpu": 101132.118428, "vram_gb": 20.709735} {"adam_lr": 3.1e-05, "dt_ms": 161.61871, "global_tokens_per_sec": 101374.401789, "grad_norm": 64.5, "loss": 5.691154, "lr_stage": "warmup", "mfu": 58.233364, "stage_name": "permutation_cycle", "status": "training", "step": 61, "tokens_per_sec_per_gpu": 101374.401789, "vram_gb": 20.709735} {"adam_lr": 3.15e-05, "dt_ms": 166.221619, "global_tokens_per_sec": 98567.202827, "grad_norm": 56.75, "loss": 5.393744, "lr_stage": "warmup", "mfu": 56.6208, "stage_name": "permutation_cycle", "status": "training", "step": 62, "tokens_per_sec_per_gpu": 98567.202827, "vram_gb": 20.709735} {"adam_lr": 3.2e-05, "dt_ms": 161.312342, "global_tokens_per_sec": 101566.934237, "grad_norm": 55.75, "loss": 5.645757, "lr_stage": "warmup", "mfu": 58.343962, "stage_name": "permutation_cycle", "status": "training", "step": 63, "tokens_per_sec_per_gpu": 101566.934237, "vram_gb": 20.709735} {"adam_lr": 3.2500000000000004e-05, "dt_ms": 160.48646, "global_tokens_per_sec": 102089.609475, "grad_norm": 61.5, "loss": 5.596967, "lr_stage": "warmup", "mfu": 58.644206, "stage_name": "permutation_cycle", "status": "training", "step": 64, "tokens_per_sec_per_gpu": 102089.609475, "vram_gb": 20.709735} {"adam_lr": 3.3e-05, "dt_ms": 159.766436, "global_tokens_per_sec": 102549.699729, "grad_norm": 58.0, "loss": 5.849578, "lr_stage": "warmup", "mfu": 58.9085, "stage_name": "permutation_cycle", "status": "training", "step": 65, "tokens_per_sec_per_gpu": 102549.699729, "vram_gb": 20.709735} {"adam_lr": 3.35e-05, "dt_ms": 161.470175, "global_tokens_per_sec": 101467.655072, "grad_norm": 61.25, "loss": 5.621767, "lr_stage": "warmup", "mfu": 58.286932, "stage_name": "permutation_cycle", "status": "training", "step": 66, "tokens_per_sec_per_gpu": 101467.655072, "vram_gb": 20.709735} {"adam_lr": 3.4e-05, "dt_ms": 163.7187, "global_tokens_per_sec": 100074.090248, "grad_norm": 52.25, "loss": 5.749467, "lr_stage": "warmup", "mfu": 57.486415, "stage_name": "permutation_cycle", "status": "training", "step": 67, "tokens_per_sec_per_gpu": 100074.090248, "vram_gb": 20.709735} {"adam_lr": 3.4500000000000005e-05, "dt_ms": 164.337397, "global_tokens_per_sec": 99697.332055, "grad_norm": 56.5, "loss": 5.893763, "lr_stage": "warmup", "mfu": 57.26999, "stage_name": "permutation_cycle", "status": "training", "step": 68, "tokens_per_sec_per_gpu": 99697.332055, "vram_gb": 20.709735} {"adam_lr": 3.5000000000000004e-05, "dt_ms": 164.096117, "global_tokens_per_sec": 99843.922559, "grad_norm": 48.25, "loss": 5.381509, "lr_stage": "warmup", "mfu": 57.354197, "stage_name": "permutation_cycle", "status": "training", "step": 69, "tokens_per_sec_per_gpu": 99843.922559, "vram_gb": 20.709735} {"adam_lr": 3.55e-05, "dt_ms": 162.862062, "global_tokens_per_sec": 100600.469828, "grad_norm": 65.0, "loss": 5.392815, "lr_stage": "warmup", "mfu": 57.788787, "stage_name": "permutation_cycle", "status": "training", "step": 70, "tokens_per_sec_per_gpu": 100600.469828, "vram_gb": 20.709735} {"adam_lr": 3.6e-05, "dt_ms": 160.450935, "global_tokens_per_sec": 102112.212452, "grad_norm": 63.75, "loss": 5.439983, "lr_stage": "warmup", "mfu": 58.65719, "stage_name": "permutation_cycle", "status": "training", "step": 71, "tokens_per_sec_per_gpu": 102112.212452, "vram_gb": 20.709735} {"adam_lr": 3.65e-05, "dt_ms": 161.447287, "global_tokens_per_sec": 101482.040017, "grad_norm": 54.0, "loss": 5.483764, "lr_stage": "warmup", "mfu": 58.295195, "stage_name": "permutation_cycle", "status": "training", "step": 72, "tokens_per_sec_per_gpu": 101482.040017, "vram_gb": 20.709735} {"adam_lr": 3.7000000000000005e-05, "dt_ms": 161.186934, "global_tokens_per_sec": 101645.956297, "grad_norm": 67.5, "loss": 5.256699, "lr_stage": "warmup", "mfu": 58.389355, "stage_name": "permutation_cycle", "status": "training", "step": 73, "tokens_per_sec_per_gpu": 101645.956297, "vram_gb": 20.709735} {"adam_lr": 3.7500000000000003e-05, "dt_ms": 163.249493, "global_tokens_per_sec": 100361.72079, "grad_norm": 55.75, "loss": 5.433264, "lr_stage": "warmup", "mfu": 57.651641, "stage_name": "permutation_cycle", "status": "training", "step": 74, "tokens_per_sec_per_gpu": 100361.72079, "vram_gb": 20.709735} {"adam_lr": 3.8e-05, "dt_ms": 163.188219, "global_tokens_per_sec": 100399.404401, "grad_norm": 58.0, "loss": 5.676508, "lr_stage": "warmup", "mfu": 57.673288, "stage_name": "permutation_cycle", "status": "training", "step": 75, "tokens_per_sec_per_gpu": 100399.404401, "vram_gb": 20.709735} {"adam_lr": 3.85e-05, "dt_ms": 163.718462, "global_tokens_per_sec": 100074.235983, "grad_norm": 53.5, "loss": 5.857623, "lr_stage": "warmup", "mfu": 57.486498, "stage_name": "permutation_cycle", "status": "training", "step": 76, "tokens_per_sec_per_gpu": 100074.235983, "vram_gb": 20.709735} {"adam_lr": 3.9e-05, "dt_ms": 161.950111, "global_tokens_per_sec": 101166.957278, "grad_norm": 64.5, "loss": 5.63308, "lr_stage": "warmup", "mfu": 58.114199, "stage_name": "permutation_cycle", "status": "training", "step": 77, "tokens_per_sec_per_gpu": 101166.957278, "vram_gb": 20.709735} {"adam_lr": 3.9500000000000005e-05, "dt_ms": 160.359144, "global_tokens_per_sec": 102170.662488, "grad_norm": 62.0, "loss": 5.337358, "lr_stage": "warmup", "mfu": 58.690766, "stage_name": "permutation_cycle", "status": "training", "step": 78, "tokens_per_sec_per_gpu": 102170.662488, "vram_gb": 20.709735} {"adam_lr": 4e-05, "dt_ms": 160.756588, "global_tokens_per_sec": 101918.062617, "grad_norm": 58.5, "loss": 5.653087, "lr_stage": "warmup", "mfu": 58.545663, "stage_name": "permutation_cycle", "status": "training", "step": 79, "tokens_per_sec_per_gpu": 101918.062617, "vram_gb": 20.709735} {"adam_lr": 4.0499999999999995e-05, "dt_ms": 163.103342, "global_tokens_per_sec": 100451.651042, "grad_norm": 61.5, "loss": 5.364152, "lr_stage": "warmup", "mfu": 57.7033, "stage_name": "permutation_cycle", "status": "training", "step": 80, "tokens_per_sec_per_gpu": 100451.651042, "vram_gb": 20.709735} {"adam_lr": 4.1e-05, "dt_ms": 162.783384, "global_tokens_per_sec": 100649.093076, "grad_norm": 58.5, "loss": 5.351438, "lr_stage": "warmup", "mfu": 57.816718, "stage_name": "permutation_cycle", "status": "training", "step": 81, "tokens_per_sec_per_gpu": 100649.093076, "vram_gb": 20.709735} {"adam_lr": 4.15e-05, "dt_ms": 161.013365, "global_tokens_per_sec": 101755.528314, "grad_norm": 61.5, "loss": 5.110683, "lr_stage": "warmup", "mfu": 58.452297, "stage_name": "permutation_cycle", "status": "training", "step": 82, "tokens_per_sec_per_gpu": 101755.528314, "vram_gb": 20.709735} {"adam_lr": 4.2000000000000004e-05, "dt_ms": 162.786245, "global_tokens_per_sec": 100647.324135, "grad_norm": 71.0, "loss": 5.426392, "lr_stage": "warmup", "mfu": 57.815702, "stage_name": "permutation_cycle", "status": "training", "step": 83, "tokens_per_sec_per_gpu": 100647.324135, "vram_gb": 20.709735} {"adam_lr": 4.25e-05, "dt_ms": 163.690329, "global_tokens_per_sec": 100091.435703, "grad_norm": 50.75, "loss": 5.769783, "lr_stage": "warmup", "mfu": 57.496378, "stage_name": "permutation_cycle", "status": "training", "step": 84, "tokens_per_sec_per_gpu": 100091.435703, "vram_gb": 20.709735} {"adam_lr": 4.3e-05, "dt_ms": 160.998821, "global_tokens_per_sec": 101764.720213, "grad_norm": 55.5, "loss": 5.364849, "lr_stage": "warmup", "mfu": 58.457578, "stage_name": "permutation_cycle", "status": "training", "step": 85, "tokens_per_sec_per_gpu": 101764.720213, "vram_gb": 20.709735} {"adam_lr": 4.350000000000001e-05, "dt_ms": 162.736177, "global_tokens_per_sec": 100678.289593, "grad_norm": 52.25, "loss": 5.267352, "lr_stage": "warmup", "mfu": 57.83349, "stage_name": "permutation_cycle", "status": "training", "step": 86, "tokens_per_sec_per_gpu": 100678.289593, "vram_gb": 20.709735} {"adam_lr": 4.4000000000000006e-05, "dt_ms": 162.823677, "global_tokens_per_sec": 100624.186209, "grad_norm": 59.25, "loss": 5.385738, "lr_stage": "warmup", "mfu": 57.802411, "stage_name": "permutation_cycle", "status": "training", "step": 87, "tokens_per_sec_per_gpu": 100624.186209, "vram_gb": 20.709735} {"adam_lr": 4.45e-05, "dt_ms": 162.445545, "global_tokens_per_sec": 100858.413693, "grad_norm": 51.75, "loss": 5.779125, "lr_stage": "warmup", "mfu": 57.93696, "stage_name": "permutation_cycle", "status": "training", "step": 88, "tokens_per_sec_per_gpu": 100858.413693, "vram_gb": 20.709735} {"adam_lr": 4.5e-05, "dt_ms": 161.47542, "global_tokens_per_sec": 101464.359097, "grad_norm": 60.0, "loss": 5.177516, "lr_stage": "warmup", "mfu": 58.285039, "stage_name": "permutation_cycle", "status": "training", "step": 89, "tokens_per_sec_per_gpu": 101464.359097, "vram_gb": 20.709735} {"adam_lr": 4.55e-05, "dt_ms": 164.422035, "global_tokens_per_sec": 99646.011426, "grad_norm": 57.5, "loss": 4.937389, "lr_stage": "warmup", "mfu": 57.24051, "stage_name": "permutation_cycle", "status": "training", "step": 90, "tokens_per_sec_per_gpu": 99646.011426, "vram_gb": 20.709735} {"adam_lr": 4.6e-05, "dt_ms": 162.889242, "global_tokens_per_sec": 100583.683622, "grad_norm": 59.75, "loss": 5.096027, "lr_stage": "warmup", "mfu": 57.779145, "stage_name": "permutation_cycle", "status": "training", "step": 91, "tokens_per_sec_per_gpu": 100583.683622, "vram_gb": 20.709735} {"adam_lr": 4.6500000000000005e-05, "dt_ms": 160.054922, "global_tokens_per_sec": 102364.861915, "grad_norm": 68.0, "loss": 5.468765, "lr_stage": "warmup", "mfu": 58.802322, "stage_name": "permutation_cycle", "status": "training", "step": 92, "tokens_per_sec_per_gpu": 102364.861915, "vram_gb": 20.709735} {"adam_lr": 4.7000000000000004e-05, "dt_ms": 163.108349, "global_tokens_per_sec": 100448.567568, "grad_norm": 60.5, "loss": 5.403374, "lr_stage": "warmup", "mfu": 57.701529, "stage_name": "permutation_cycle", "status": "training", "step": 93, "tokens_per_sec_per_gpu": 100448.567568, "vram_gb": 20.709735} {"adam_lr": 4.7499999999999996e-05, "dt_ms": 162.460089, "global_tokens_per_sec": 100849.384782, "grad_norm": 52.0, "loss": 5.16357, "lr_stage": "warmup", "mfu": 57.931774, "stage_name": "permutation_cycle", "status": "training", "step": 94, "tokens_per_sec_per_gpu": 100849.384782, "vram_gb": 20.709735} {"adam_lr": 4.8e-05, "dt_ms": 163.211584, "global_tokens_per_sec": 100385.031438, "grad_norm": 54.5, "loss": 5.492831, "lr_stage": "warmup", "mfu": 57.665031, "stage_name": "permutation_cycle", "status": "training", "step": 95, "tokens_per_sec_per_gpu": 100385.031438, "vram_gb": 20.709735} {"adam_lr": 4.85e-05, "dt_ms": 161.669016, "global_tokens_per_sec": 101342.857259, "grad_norm": 67.0, "loss": 5.075224, "lr_stage": "warmup", "mfu": 58.215243, "stage_name": "permutation_cycle", "status": "training", "step": 96, "tokens_per_sec_per_gpu": 101342.857259, "vram_gb": 20.709735} {"adam_lr": 4.9e-05, "dt_ms": 164.019108, "global_tokens_per_sec": 99890.800638, "grad_norm": 55.0, "loss": 4.906778, "lr_stage": "warmup", "mfu": 57.381126, "stage_name": "permutation_cycle", "status": "training", "step": 97, "tokens_per_sec_per_gpu": 99890.800638, "vram_gb": 20.709735} {"adam_lr": 4.9500000000000004e-05, "dt_ms": 164.330721, "global_tokens_per_sec": 99701.382128, "grad_norm": 50.0, "loss": 5.692781, "lr_stage": "warmup", "mfu": 57.272317, "stage_name": "permutation_cycle", "status": "training", "step": 98, "tokens_per_sec_per_gpu": 99701.382128, "vram_gb": 20.709735} {"adam_lr": 5e-05, "dt_ms": 164.280653, "global_tokens_per_sec": 99731.768171, "grad_norm": 50.25, "loss": 5.049854, "lr_stage": "warmup", "mfu": 57.289772, "stage_name": "permutation_cycle", "status": "training", "step": 99, "tokens_per_sec_per_gpu": 99731.768171, "vram_gb": 20.709735} {"adam_lr": 5.050000000000001e-05, "dt_ms": 161.717892, "global_tokens_per_sec": 101312.228526, "grad_norm": 64.5, "loss": 4.700605, "lr_stage": "warmup", "mfu": 58.197649, "stage_name": "permutation_cycle", "status": "training", "step": 100, "tokens_per_sec_per_gpu": 101312.228526, "vram_gb": 20.709735} {"adam_lr": 5.1000000000000006e-05, "dt_ms": 163.663387, "global_tokens_per_sec": 100107.912163, "grad_norm": 60.25, "loss": 5.138827, "lr_stage": "warmup", "mfu": 57.505843, "stage_name": "permutation_cycle", "status": "training", "step": 101, "tokens_per_sec_per_gpu": 100107.912163, "vram_gb": 20.709735} {"adam_lr": 5.15e-05, "dt_ms": 161.945581, "global_tokens_per_sec": 101169.787127, "grad_norm": 62.25, "loss": 4.589236, "lr_stage": "warmup", "mfu": 58.115825, "stage_name": "permutation_cycle", "status": "training", "step": 102, "tokens_per_sec_per_gpu": 101169.787127, "vram_gb": 20.709735} {"adam_lr": 5.2000000000000004e-05, "dt_ms": 164.928198, "global_tokens_per_sec": 99340.199023, "grad_norm": 55.5, "loss": 4.859569, "lr_stage": "warmup", "mfu": 57.064839, "stage_name": "permutation_cycle", "status": "training", "step": 103, "tokens_per_sec_per_gpu": 99340.199023, "vram_gb": 20.709735} {"adam_lr": 5.25e-05, "dt_ms": 163.804531, "global_tokens_per_sec": 100021.653188, "grad_norm": 59.0, "loss": 4.952992, "lr_stage": "warmup", "mfu": 57.456293, "stage_name": "permutation_cycle", "status": "training", "step": 104, "tokens_per_sec_per_gpu": 100021.653188, "vram_gb": 20.709735} {"adam_lr": 5.3e-05, "dt_ms": 164.158344, "global_tokens_per_sec": 99806.074878, "grad_norm": 57.0, "loss": 4.878649, "lr_stage": "warmup", "mfu": 57.332456, "stage_name": "permutation_cycle", "status": "training", "step": 105, "tokens_per_sec_per_gpu": 99806.074878, "vram_gb": 20.709735} {"adam_lr": 5.3500000000000006e-05, "dt_ms": 164.920092, "global_tokens_per_sec": 99345.08184, "grad_norm": 49.0, "loss": 5.125963, "lr_stage": "warmup", "mfu": 57.067644, "stage_name": "permutation_cycle", "status": "training", "step": 106, "tokens_per_sec_per_gpu": 99345.08184, "vram_gb": 20.709735} {"adam_lr": 5.4000000000000005e-05, "dt_ms": 164.481878, "global_tokens_per_sec": 99609.757447, "grad_norm": 57.0, "loss": 5.111421, "lr_stage": "warmup", "mfu": 57.219684, "stage_name": "permutation_cycle", "status": "training", "step": 107, "tokens_per_sec_per_gpu": 99609.757447, "vram_gb": 20.709735} {"adam_lr": 5.4499999999999997e-05, "dt_ms": 162.80365, "global_tokens_per_sec": 100636.564413, "grad_norm": 55.5, "loss": 4.666306, "lr_stage": "warmup", "mfu": 57.809521, "stage_name": "permutation_cycle", "status": "training", "step": 108, "tokens_per_sec_per_gpu": 100636.564413, "vram_gb": 20.709735} {"adam_lr": 5.500000000000001e-05, "dt_ms": 166.054726, "global_tokens_per_sec": 98666.267618, "grad_norm": 61.75, "loss": 5.343162, "lr_stage": "warmup", "mfu": 56.677707, "stage_name": "permutation_cycle", "status": "training", "step": 109, "tokens_per_sec_per_gpu": 98666.267618, "vram_gb": 20.709735} {"adam_lr": 5.55e-05, "dt_ms": 163.517237, "global_tokens_per_sec": 100197.387931, "grad_norm": 58.0, "loss": 3.979779, "lr_stage": "warmup", "mfu": 57.557241, "stage_name": "permutation_cycle", "status": "training", "step": 110, "tokens_per_sec_per_gpu": 100197.387931, "vram_gb": 20.709735} {"adam_lr": 5.6e-05, "dt_ms": 163.741112, "global_tokens_per_sec": 100060.393046, "grad_norm": 55.75, "loss": 5.444377, "lr_stage": "warmup", "mfu": 57.478546, "stage_name": "permutation_cycle", "status": "training", "step": 111, "tokens_per_sec_per_gpu": 100060.393046, "vram_gb": 20.709735} {"adam_lr": 5.6500000000000005e-05, "dt_ms": 163.575172, "global_tokens_per_sec": 100161.899616, "grad_norm": 51.5, "loss": 4.865255, "lr_stage": "warmup", "mfu": 57.536856, "stage_name": "permutation_cycle", "status": "training", "step": 112, "tokens_per_sec_per_gpu": 100161.899616, "vram_gb": 20.709735} {"adam_lr": 5.7e-05, "dt_ms": 165.428638, "global_tokens_per_sec": 99039.683532, "grad_norm": 54.25, "loss": 5.14979, "lr_stage": "warmup", "mfu": 56.892211, "stage_name": "permutation_cycle", "status": "training", "step": 113, "tokens_per_sec_per_gpu": 99039.683532, "vram_gb": 20.709735} {"adam_lr": 5.75e-05, "dt_ms": 162.399292, "global_tokens_per_sec": 100887.13934, "grad_norm": 65.0, "loss": 4.601436, "lr_stage": "warmup", "mfu": 57.953461, "stage_name": "permutation_cycle", "status": "training", "step": 114, "tokens_per_sec_per_gpu": 100887.13934, "vram_gb": 20.709735} {"adam_lr": 5.800000000000001e-05, "dt_ms": 165.40575, "global_tokens_per_sec": 99053.388246, "grad_norm": 59.0, "loss": 4.831189, "lr_stage": "warmup", "mfu": 56.900084, "stage_name": "permutation_cycle", "status": "training", "step": 115, "tokens_per_sec_per_gpu": 99053.388246, "vram_gb": 20.709735} {"adam_lr": 5.85e-05, "dt_ms": 163.635969, "global_tokens_per_sec": 100124.685813, "grad_norm": 57.5, "loss": 4.92204, "lr_stage": "warmup", "mfu": 57.515479, "stage_name": "permutation_cycle", "status": "training", "step": 116, "tokens_per_sec_per_gpu": 100124.685813, "vram_gb": 20.709735} {"adam_lr": 5.9e-05, "dt_ms": 161.665678, "global_tokens_per_sec": 101344.949653, "grad_norm": 62.25, "loss": 4.136198, "lr_stage": "warmup", "mfu": 58.216445, "stage_name": "permutation_cycle", "status": "training", "step": 117, "tokens_per_sec_per_gpu": 101344.949653, "vram_gb": 20.709735} {"adam_lr": 5.95e-05, "dt_ms": 162.936926, "global_tokens_per_sec": 100554.247668, "grad_norm": 57.5, "loss": 4.705051, "lr_stage": "warmup", "mfu": 57.762235, "stage_name": "permutation_cycle", "status": "training", "step": 118, "tokens_per_sec_per_gpu": 100554.247668, "vram_gb": 20.709735} {"adam_lr": 6e-05, "dt_ms": 163.484335, "global_tokens_per_sec": 100217.552987, "grad_norm": 56.25, "loss": 4.443789, "lr_stage": "warmup", "mfu": 57.568825, "stage_name": "permutation_cycle", "status": "training", "step": 119, "tokens_per_sec_per_gpu": 100217.552987, "vram_gb": 20.709735} {"adam_lr": 6.050000000000001e-05, "dt_ms": 163.230658, "global_tokens_per_sec": 100373.301457, "grad_norm": 53.5, "loss": 4.662738, "lr_stage": "warmup", "mfu": 57.658293, "stage_name": "permutation_cycle", "status": "training", "step": 120, "tokens_per_sec_per_gpu": 100373.301457, "vram_gb": 20.709735} {"adam_lr": 6.1000000000000005e-05, "dt_ms": 165.577888, "global_tokens_per_sec": 98950.410284, "grad_norm": 53.75, "loss": 4.843942, "lr_stage": "warmup", "mfu": 56.840929, "stage_name": "permutation_cycle", "status": "training", "step": 121, "tokens_per_sec_per_gpu": 98950.410284, "vram_gb": 20.709735} {"adam_lr": 6.15e-05, "dt_ms": 163.865328, "global_tokens_per_sec": 99984.543506, "grad_norm": 57.75, "loss": 4.485537, "lr_stage": "warmup", "mfu": 57.434975, "stage_name": "permutation_cycle", "status": "training", "step": 122, "tokens_per_sec_per_gpu": 99984.543506, "vram_gb": 20.709735} {"adam_lr": 6.2e-05, "dt_ms": 162.829161, "global_tokens_per_sec": 100620.79747, "grad_norm": 51.75, "loss": 4.402995, "lr_stage": "warmup", "mfu": 57.800464, "stage_name": "permutation_cycle", "status": "training", "step": 123, "tokens_per_sec_per_gpu": 100620.79747, "vram_gb": 20.709735} {"adam_lr": 6.25e-05, "dt_ms": 164.22534, "global_tokens_per_sec": 99765.359055, "grad_norm": 53.25, "loss": 4.854642, "lr_stage": "warmup", "mfu": 57.309067, "stage_name": "permutation_cycle", "status": "training", "step": 124, "tokens_per_sec_per_gpu": 99765.359055, "vram_gb": 20.709735} {"adam_lr": 6.3e-05, "dt_ms": 163.256645, "global_tokens_per_sec": 100357.323769, "grad_norm": 50.25, "loss": 4.868283, "lr_stage": "warmup", "mfu": 57.649115, "stage_name": "permutation_cycle", "status": "training", "step": 125, "tokens_per_sec_per_gpu": 100357.323769, "vram_gb": 20.709735} {"adam_lr": 6.350000000000001e-05, "dt_ms": 161.492586, "global_tokens_per_sec": 101453.573765, "grad_norm": 62.0, "loss": 3.96059, "lr_stage": "warmup", "mfu": 58.278843, "stage_name": "permutation_cycle", "status": "training", "step": 126, "tokens_per_sec_per_gpu": 101453.573765, "vram_gb": 20.709735} {"adam_lr": 6.4e-05, "dt_ms": 163.950443, "global_tokens_per_sec": 99932.636188, "grad_norm": 51.5, "loss": 3.970458, "lr_stage": "warmup", "mfu": 57.405158, "stage_name": "permutation_cycle", "status": "training", "step": 127, "tokens_per_sec_per_gpu": 99932.636188, "vram_gb": 20.709735} {"adam_lr": 6.45e-05, "dt_ms": 163.161516, "global_tokens_per_sec": 100415.835686, "grad_norm": 54.0, "loss": 3.609528, "lr_stage": "warmup", "mfu": 57.682726, "stage_name": "permutation_cycle", "status": "training", "step": 128, "tokens_per_sec_per_gpu": 100415.835686, "vram_gb": 20.709735} {"adam_lr": 6.500000000000001e-05, "dt_ms": 161.738157, "global_tokens_per_sec": 101299.534237, "grad_norm": 55.25, "loss": 4.070619, "lr_stage": "warmup", "mfu": 58.190357, "stage_name": "permutation_cycle", "status": "training", "step": 129, "tokens_per_sec_per_gpu": 101299.534237, "vram_gb": 20.709735} {"adam_lr": 6.55e-05, "dt_ms": 163.30266, "global_tokens_per_sec": 100329.045474, "grad_norm": 53.75, "loss": 4.477702, "lr_stage": "warmup", "mfu": 57.632871, "stage_name": "permutation_cycle", "status": "training", "step": 130, "tokens_per_sec_per_gpu": 100329.045474, "vram_gb": 20.709735} {"adam_lr": 6.6e-05, "dt_ms": 164.074421, "global_tokens_per_sec": 99857.125244, "grad_norm": 55.75, "loss": 4.790458, "lr_stage": "warmup", "mfu": 57.361781, "stage_name": "permutation_cycle", "status": "training", "step": 131, "tokens_per_sec_per_gpu": 99857.125244, "vram_gb": 20.709735} {"adam_lr": 6.65e-05, "dt_ms": 163.175106, "global_tokens_per_sec": 100407.472664, "grad_norm": 50.5, "loss": 4.112131, "lr_stage": "warmup", "mfu": 57.677922, "stage_name": "permutation_cycle", "status": "training", "step": 132, "tokens_per_sec_per_gpu": 100407.472664, "vram_gb": 20.709735} {"adam_lr": 6.7e-05, "dt_ms": 163.128376, "global_tokens_per_sec": 100436.235565, "grad_norm": 51.0, "loss": 4.607268, "lr_stage": "warmup", "mfu": 57.694445, "stage_name": "permutation_cycle", "status": "training", "step": 133, "tokens_per_sec_per_gpu": 100436.235565, "vram_gb": 20.709735} {"adam_lr": 6.75e-05, "dt_ms": 165.246964, "global_tokens_per_sec": 99148.569226, "grad_norm": 53.0, "loss": 4.56246, "lr_stage": "warmup", "mfu": 56.95476, "stage_name": "permutation_cycle", "status": "training", "step": 134, "tokens_per_sec_per_gpu": 99148.569226, "vram_gb": 20.709735} {"adam_lr": 6.8e-05, "dt_ms": 162.080288, "global_tokens_per_sec": 101085.703937, "grad_norm": 55.0, "loss": 4.462828, "lr_stage": "warmup", "mfu": 58.067524, "stage_name": "permutation_cycle", "status": "training", "step": 135, "tokens_per_sec_per_gpu": 101085.703937, "vram_gb": 20.709735} {"adam_lr": 6.85e-05, "dt_ms": 162.340879, "global_tokens_per_sec": 100923.439965, "grad_norm": 56.75, "loss": 3.941256, "lr_stage": "warmup", "mfu": 57.974314, "stage_name": "permutation_cycle", "status": "training", "step": 136, "tokens_per_sec_per_gpu": 100923.439965, "vram_gb": 20.709735} {"adam_lr": 6.900000000000001e-05, "dt_ms": 161.699772, "global_tokens_per_sec": 101323.581409, "grad_norm": 48.0, "loss": 4.602213, "lr_stage": "warmup", "mfu": 58.20417, "stage_name": "permutation_cycle", "status": "training", "step": 137, "tokens_per_sec_per_gpu": 101323.581409, "vram_gb": 20.709735} {"adam_lr": 6.950000000000001e-05, "dt_ms": 163.285494, "global_tokens_per_sec": 100339.593026, "grad_norm": 53.0, "loss": 4.600029, "lr_stage": "warmup", "mfu": 57.63893, "stage_name": "permutation_cycle", "status": "training", "step": 138, "tokens_per_sec_per_gpu": 100339.593026, "vram_gb": 20.709735} {"adam_lr": 7.000000000000001e-05, "dt_ms": 165.189981, "global_tokens_per_sec": 99182.770378, "grad_norm": 47.75, "loss": 4.010683, "lr_stage": "warmup", "mfu": 56.974406, "stage_name": "permutation_cycle", "status": "training", "step": 139, "tokens_per_sec_per_gpu": 99182.770378, "vram_gb": 20.709735} {"adam_lr": 7.05e-05, "dt_ms": 164.201975, "global_tokens_per_sec": 99779.555106, "grad_norm": 46.75, "loss": 3.817906, "lr_stage": "warmup", "mfu": 57.317222, "stage_name": "permutation_cycle", "status": "training", "step": 140, "tokens_per_sec_per_gpu": 99779.555106, "vram_gb": 20.709735} {"adam_lr": 7.1e-05, "dt_ms": 163.032293, "global_tokens_per_sec": 100495.427417, "grad_norm": 53.75, "loss": 4.769811, "lr_stage": "warmup", "mfu": 57.728447, "stage_name": "permutation_cycle", "status": "training", "step": 141, "tokens_per_sec_per_gpu": 100495.427417, "vram_gb": 20.709735} {"adam_lr": 7.15e-05, "dt_ms": 163.86795, "global_tokens_per_sec": 99982.943315, "grad_norm": 47.0, "loss": 3.636345, "lr_stage": "warmup", "mfu": 57.434056, "stage_name": "permutation_cycle", "status": "training", "step": 142, "tokens_per_sec_per_gpu": 99982.943315, "vram_gb": 20.709735} {"adam_lr": 7.2e-05, "dt_ms": 162.90307, "global_tokens_per_sec": 100575.145421, "grad_norm": 50.25, "loss": 4.102437, "lr_stage": "warmup", "mfu": 57.77424, "stage_name": "permutation_cycle", "status": "training", "step": 143, "tokens_per_sec_per_gpu": 100575.145421, "vram_gb": 20.709735} {"adam_lr": 7.25e-05, "dt_ms": 161.987782, "global_tokens_per_sec": 101143.430979, "grad_norm": 51.0, "loss": 4.029489, "lr_stage": "warmup", "mfu": 58.100685, "stage_name": "permutation_cycle", "status": "training", "step": 144, "tokens_per_sec_per_gpu": 101143.430979, "vram_gb": 20.709735} {"adam_lr": 7.3e-05, "dt_ms": 163.2092, "global_tokens_per_sec": 100386.497878, "grad_norm": 56.25, "loss": 4.115476, "lr_stage": "warmup", "mfu": 57.665874, "stage_name": "permutation_cycle", "status": "training", "step": 145, "tokens_per_sec_per_gpu": 100386.497878, "vram_gb": 20.709735} {"adam_lr": 7.350000000000001e-05, "dt_ms": 163.892508, "global_tokens_per_sec": 99967.9622, "grad_norm": 53.75, "loss": 4.155134, "lr_stage": "warmup", "mfu": 57.42545, "stage_name": "permutation_cycle", "status": "training", "step": 146, "tokens_per_sec_per_gpu": 99967.9622, "vram_gb": 20.709735} {"adam_lr": 7.400000000000001e-05, "dt_ms": 162.06646, "global_tokens_per_sec": 101094.329047, "grad_norm": 50.25, "loss": 3.947436, "lr_stage": "warmup", "mfu": 58.072479, "stage_name": "permutation_cycle", "status": "training", "step": 147, "tokens_per_sec_per_gpu": 101094.329047, "vram_gb": 20.709735} {"adam_lr": 7.45e-05, "dt_ms": 162.303925, "global_tokens_per_sec": 100946.41916, "grad_norm": 57.5, "loss": 3.084353, "lr_stage": "warmup", "mfu": 57.987514, "stage_name": "permutation_cycle", "status": "training", "step": 148, "tokens_per_sec_per_gpu": 100946.41916, "vram_gb": 20.709735} {"adam_lr": 7.500000000000001e-05, "dt_ms": 162.811756, "global_tokens_per_sec": 100631.55382, "grad_norm": 52.75, "loss": 4.170026, "lr_stage": "warmup", "mfu": 57.806643, "stage_name": "permutation_cycle", "status": "training", "step": 149, "tokens_per_sec_per_gpu": 100631.55382, "vram_gb": 20.709735} {"adam_lr": 7.55e-05, "dt_ms": 162.580013, "global_tokens_per_sec": 100774.994847, "grad_norm": 45.0, "loss": 3.319028, "lr_stage": "warmup", "mfu": 57.889041, "stage_name": "permutation_cycle", "status": "training", "step": 150, "tokens_per_sec_per_gpu": 100774.994847, "vram_gb": 20.709735} {"adam_lr": 7.6e-05, "dt_ms": 163.505077, "global_tokens_per_sec": 100204.839289, "grad_norm": 50.25, "loss": 3.973546, "lr_stage": "warmup", "mfu": 57.561522, "stage_name": "permutation_cycle", "status": "training", "step": 151, "tokens_per_sec_per_gpu": 100204.839289, "vram_gb": 20.709735} {"adam_lr": 7.65e-05, "dt_ms": 163.011789, "global_tokens_per_sec": 100508.067963, "grad_norm": 44.0, "loss": 4.207307, "lr_stage": "warmup", "mfu": 57.735708, "stage_name": "permutation_cycle", "status": "training", "step": 152, "tokens_per_sec_per_gpu": 100508.067963, "vram_gb": 20.709735} {"adam_lr": 7.7e-05, "dt_ms": 160.946369, "global_tokens_per_sec": 101797.885124, "grad_norm": 47.25, "loss": 3.90346, "lr_stage": "warmup", "mfu": 58.476629, "stage_name": "permutation_cycle", "status": "training", "step": 153, "tokens_per_sec_per_gpu": 101797.885124, "vram_gb": 20.709735} {"adam_lr": 7.75e-05, "dt_ms": 162.597179, "global_tokens_per_sec": 100764.355564, "grad_norm": 51.0, "loss": 3.964319, "lr_stage": "warmup", "mfu": 57.88293, "stage_name": "permutation_cycle", "status": "training", "step": 154, "tokens_per_sec_per_gpu": 100764.355564, "vram_gb": 20.709735} {"adam_lr": 7.8e-05, "dt_ms": 162.230968, "global_tokens_per_sec": 100991.815274, "grad_norm": 53.25, "loss": 4.047668, "lr_stage": "warmup", "mfu": 58.013591, "stage_name": "permutation_cycle", "status": "training", "step": 155, "tokens_per_sec_per_gpu": 100991.815274, "vram_gb": 20.709735} {"adam_lr": 7.850000000000001e-05, "dt_ms": 163.440704, "global_tokens_per_sec": 100244.306127, "grad_norm": 46.25, "loss": 3.702322, "lr_stage": "warmup", "mfu": 57.584193, "stage_name": "permutation_cycle", "status": "training", "step": 156, "tokens_per_sec_per_gpu": 100244.306127, "vram_gb": 20.709735} {"adam_lr": 7.900000000000001e-05, "dt_ms": 162.976742, "global_tokens_per_sec": 100529.681843, "grad_norm": 55.0, "loss": 3.447516, "lr_stage": "warmup", "mfu": 57.748124, "stage_name": "permutation_cycle", "status": "training", "step": 157, "tokens_per_sec_per_gpu": 100529.681843, "vram_gb": 20.709735} {"adam_lr": 7.950000000000001e-05, "dt_ms": 163.273573, "global_tokens_per_sec": 100346.919019, "grad_norm": 47.25, "loss": 2.911187, "lr_stage": "warmup", "mfu": 57.643138, "stage_name": "permutation_cycle", "status": "training", "step": 158, "tokens_per_sec_per_gpu": 100346.919019, "vram_gb": 20.709735} {"adam_lr": 8e-05, "dt_ms": 162.33325, "global_tokens_per_sec": 100928.1832, "grad_norm": 47.25, "loss": 3.493585, "lr_stage": "warmup", "mfu": 57.977038, "stage_name": "permutation_cycle", "status": "training", "step": 159, "tokens_per_sec_per_gpu": 100928.1832, "vram_gb": 20.709735} {"adam_lr": 8.05e-05, "dt_ms": 162.287951, "global_tokens_per_sec": 100956.355342, "grad_norm": 42.25, "loss": 3.666064, "lr_stage": "warmup", "mfu": 57.993222, "stage_name": "permutation_cycle", "status": "training", "step": 160, "tokens_per_sec_per_gpu": 100956.355342, "vram_gb": 20.709735} {"adam_lr": 8.099999999999999e-05, "dt_ms": 161.915541, "global_tokens_per_sec": 101188.557501, "grad_norm": 39.75, "loss": 3.183924, "lr_stage": "warmup", "mfu": 58.126607, "stage_name": "permutation_cycle", "status": "training", "step": 161, "tokens_per_sec_per_gpu": 101188.557501, "vram_gb": 20.709735} {"adam_lr": 8.150000000000002e-05, "dt_ms": 162.166119, "global_tokens_per_sec": 101032.201666, "grad_norm": 42.5, "loss": 3.145398, "lr_stage": "warmup", "mfu": 58.036791, "stage_name": "permutation_cycle", "status": "training", "step": 162, "tokens_per_sec_per_gpu": 101032.201666, "vram_gb": 20.709735} {"adam_lr": 8.2e-05, "dt_ms": 163.949728, "global_tokens_per_sec": 99933.07216, "grad_norm": 48.5, "loss": 3.151726, "lr_stage": "warmup", "mfu": 57.405408, "stage_name": "permutation_cycle", "status": "training", "step": 163, "tokens_per_sec_per_gpu": 99933.07216, "vram_gb": 20.709735} {"adam_lr": 8.25e-05, "dt_ms": 162.834406, "global_tokens_per_sec": 100617.556281, "grad_norm": 43.25, "loss": 3.405162, "lr_stage": "warmup", "mfu": 57.798602, "stage_name": "permutation_cycle", "status": "training", "step": 164, "tokens_per_sec_per_gpu": 100617.556281, "vram_gb": 20.709735} {"adam_lr": 8.3e-05, "dt_ms": 161.886215, "global_tokens_per_sec": 101206.887682, "grad_norm": 48.25, "loss": 2.924721, "lr_stage": "warmup", "mfu": 58.137137, "stage_name": "permutation_cycle", "status": "training", "step": 165, "tokens_per_sec_per_gpu": 101206.887682, "vram_gb": 20.709735} {"adam_lr": 8.35e-05, "dt_ms": 163.273811, "global_tokens_per_sec": 100346.772489, "grad_norm": 44.0, "loss": 4.156876, "lr_stage": "warmup", "mfu": 57.643054, "stage_name": "permutation_cycle", "status": "training", "step": 166, "tokens_per_sec_per_gpu": 100346.772489, "vram_gb": 20.709735} {"adam_lr": 8.400000000000001e-05, "dt_ms": 161.775589, "global_tokens_per_sec": 101276.0955, "grad_norm": 57.25, "loss": 3.895454, "lr_stage": "warmup", "mfu": 58.176893, "stage_name": "permutation_cycle", "status": "training", "step": 167, "tokens_per_sec_per_gpu": 101276.0955, "vram_gb": 20.709735} {"adam_lr": 8.450000000000001e-05, "dt_ms": 165.349007, "global_tokens_per_sec": 99087.380878, "grad_norm": 39.75, "loss": 3.49825, "lr_stage": "warmup", "mfu": 56.919611, "stage_name": "permutation_cycle", "status": "training", "step": 168, "tokens_per_sec_per_gpu": 99087.380878, "vram_gb": 20.709735} {"adam_lr": 8.5e-05, "dt_ms": 161.634684, "global_tokens_per_sec": 101364.383152, "grad_norm": 53.25, "loss": 2.659504, "lr_stage": "warmup", "mfu": 58.227608, "stage_name": "permutation_cycle", "status": "training", "step": 169, "tokens_per_sec_per_gpu": 101364.383152, "vram_gb": 20.709735} {"adam_lr": 8.55e-05, "dt_ms": 160.863161, "global_tokens_per_sec": 101850.541101, "grad_norm": 47.25, "loss": 3.299463, "lr_stage": "warmup", "mfu": 58.506876, "stage_name": "permutation_cycle", "status": "training", "step": 170, "tokens_per_sec_per_gpu": 101850.541101, "vram_gb": 20.709735} {"adam_lr": 8.6e-05, "dt_ms": 165.264368, "global_tokens_per_sec": 99138.12755, "grad_norm": 46.25, "loss": 3.486399, "lr_stage": "warmup", "mfu": 56.948761, "stage_name": "permutation_cycle", "status": "training", "step": 171, "tokens_per_sec_per_gpu": 99138.12755, "vram_gb": 20.709735} {"adam_lr": 8.65e-05, "dt_ms": 162.749529, "global_tokens_per_sec": 100670.030274, "grad_norm": 45.25, "loss": 3.618549, "lr_stage": "warmup", "mfu": 57.828745, "stage_name": "permutation_cycle", "status": "training", "step": 172, "tokens_per_sec_per_gpu": 100670.030274, "vram_gb": 20.709735} {"adam_lr": 8.700000000000001e-05, "dt_ms": 164.339781, "global_tokens_per_sec": 99695.88568, "grad_norm": 40.25, "loss": 3.879999, "lr_stage": "warmup", "mfu": 57.269159, "stage_name": "permutation_cycle", "status": "training", "step": 173, "tokens_per_sec_per_gpu": 99695.88568, "vram_gb": 20.709735} {"adam_lr": 8.750000000000001e-05, "dt_ms": 162.611485, "global_tokens_per_sec": 100755.49121, "grad_norm": 45.5, "loss": 3.863403, "lr_stage": "warmup", "mfu": 57.877837, "stage_name": "permutation_cycle", "status": "training", "step": 174, "tokens_per_sec_per_gpu": 100755.49121, "vram_gb": 20.709735} {"adam_lr": 8.800000000000001e-05, "dt_ms": 163.000822, "global_tokens_per_sec": 100514.830491, "grad_norm": 37.75, "loss": 3.083014, "lr_stage": "warmup", "mfu": 57.739593, "stage_name": "permutation_cycle", "status": "training", "step": 175, "tokens_per_sec_per_gpu": 100514.830491, "vram_gb": 20.709735} {"adam_lr": 8.85e-05, "dt_ms": 163.625956, "global_tokens_per_sec": 100130.813243, "grad_norm": 50.0, "loss": 3.129879, "lr_stage": "warmup", "mfu": 57.518998, "stage_name": "permutation_cycle", "status": "training", "step": 176, "tokens_per_sec_per_gpu": 100130.813243, "vram_gb": 20.709735} {"adam_lr": 8.9e-05, "dt_ms": 163.273573, "global_tokens_per_sec": 100346.919019, "grad_norm": 44.5, "loss": 2.872918, "lr_stage": "warmup", "mfu": 57.643138, "stage_name": "permutation_cycle", "status": "training", "step": 177, "tokens_per_sec_per_gpu": 100346.919019, "vram_gb": 20.709735} {"adam_lr": 8.95e-05, "dt_ms": 164.093971, "global_tokens_per_sec": 99845.228164, "grad_norm": 43.75, "loss": 3.887512, "lr_stage": "warmup", "mfu": 57.354947, "stage_name": "permutation_cycle", "status": "training", "step": 178, "tokens_per_sec_per_gpu": 99845.228164, "vram_gb": 20.709735} {"adam_lr": 9e-05, "dt_ms": 163.334131, "global_tokens_per_sec": 100309.714054, "grad_norm": 40.0, "loss": 3.206604, "lr_stage": "warmup", "mfu": 57.621766, "stage_name": "permutation_cycle", "status": "training", "step": 179, "tokens_per_sec_per_gpu": 100309.714054, "vram_gb": 20.709735} {"adam_lr": 9.05e-05, "dt_ms": 163.777351, "global_tokens_per_sec": 100038.25231, "grad_norm": 50.25, "loss": 3.897879, "lr_stage": "warmup", "mfu": 57.465828, "stage_name": "permutation_cycle", "status": "training", "step": 180, "tokens_per_sec_per_gpu": 100038.25231, "vram_gb": 20.709735} {"adam_lr": 9.1e-05, "dt_ms": 163.694143, "global_tokens_per_sec": 100089.103191, "grad_norm": 52.25, "loss": 3.750837, "lr_stage": "warmup", "mfu": 57.495039, "stage_name": "permutation_cycle", "status": "training", "step": 181, "tokens_per_sec_per_gpu": 100089.103191, "vram_gb": 20.709735} {"adam_lr": 9.15e-05, "dt_ms": 162.178278, "global_tokens_per_sec": 101024.626757, "grad_norm": 38.75, "loss": 3.494901, "lr_stage": "warmup", "mfu": 58.032439, "stage_name": "permutation_cycle", "status": "training", "step": 182, "tokens_per_sec_per_gpu": 101024.626757, "vram_gb": 20.709735} {"adam_lr": 9.2e-05, "dt_ms": 162.287474, "global_tokens_per_sec": 100956.651975, "grad_norm": 54.0, "loss": 3.418139, "lr_stage": "warmup", "mfu": 57.993392, "stage_name": "permutation_cycle", "status": "training", "step": 183, "tokens_per_sec_per_gpu": 100956.651975, "vram_gb": 20.709735} {"adam_lr": 9.250000000000001e-05, "dt_ms": 164.38961, "global_tokens_per_sec": 99665.666042, "grad_norm": 44.0, "loss": 2.77391, "lr_stage": "warmup", "mfu": 57.2518, "stage_name": "permutation_cycle", "status": "training", "step": 184, "tokens_per_sec_per_gpu": 99665.666042, "vram_gb": 20.709735} {"adam_lr": 9.300000000000001e-05, "dt_ms": 163.193464, "global_tokens_per_sec": 100396.177459, "grad_norm": 42.5, "loss": 2.722806, "lr_stage": "warmup", "mfu": 57.671434, "stage_name": "permutation_cycle", "status": "training", "step": 185, "tokens_per_sec_per_gpu": 100396.177459, "vram_gb": 20.709735} {"adam_lr": 9.350000000000001e-05, "dt_ms": 162.379265, "global_tokens_per_sec": 100899.582327, "grad_norm": 38.5, "loss": 2.900477, "lr_stage": "warmup", "mfu": 57.960609, "stage_name": "permutation_cycle", "status": "training", "step": 186, "tokens_per_sec_per_gpu": 100899.582327, "vram_gb": 20.709735} {"adam_lr": 9.400000000000001e-05, "dt_ms": 164.839506, "global_tokens_per_sec": 99393.648906, "grad_norm": 47.75, "loss": 3.224312, "lr_stage": "warmup", "mfu": 57.095543, "stage_name": "permutation_cycle", "status": "training", "step": 187, "tokens_per_sec_per_gpu": 99393.648906, "vram_gb": 20.709735} {"adam_lr": 9.45e-05, "dt_ms": 162.844658, "global_tokens_per_sec": 100611.221832, "grad_norm": 39.5, "loss": 3.110501, "lr_stage": "warmup", "mfu": 57.794964, "stage_name": "permutation_cycle", "status": "training", "step": 188, "tokens_per_sec_per_gpu": 100611.221832, "vram_gb": 20.709735} {"adam_lr": 9.499999999999999e-05, "dt_ms": 161.298752, "global_tokens_per_sec": 101575.491527, "grad_norm": 48.25, "loss": 4.069315, "lr_stage": "warmup", "mfu": 58.348877, "stage_name": "permutation_cycle", "status": "training", "step": 189, "tokens_per_sec_per_gpu": 101575.491527, "vram_gb": 20.709735} {"adam_lr": 9.550000000000002e-05, "dt_ms": 162.185907, "global_tokens_per_sec": 101019.874453, "grad_norm": 44.0, "loss": 3.852484, "lr_stage": "warmup", "mfu": 58.029709, "stage_name": "permutation_cycle", "status": "training", "step": 190, "tokens_per_sec_per_gpu": 101019.874453, "vram_gb": 20.709735} {"adam_lr": 9.6e-05, "dt_ms": 163.86795, "global_tokens_per_sec": 99982.943315, "grad_norm": 51.5, "loss": 3.568018, "lr_stage": "warmup", "mfu": 57.434056, "stage_name": "permutation_cycle", "status": "training", "step": 191, "tokens_per_sec_per_gpu": 99982.943315, "vram_gb": 20.709735} {"adam_lr": 9.65e-05, "dt_ms": 164.376974, "global_tokens_per_sec": 99673.327661, "grad_norm": 51.0, "loss": 3.299252, "lr_stage": "warmup", "mfu": 57.256201, "stage_name": "permutation_cycle", "status": "training", "step": 192, "tokens_per_sec_per_gpu": 99673.327661, "vram_gb": 20.709735} {"adam_lr": 9.7e-05, "dt_ms": 165.473938, "global_tokens_per_sec": 99012.570796, "grad_norm": 43.0, "loss": 2.868783, "lr_stage": "warmup", "mfu": 56.876637, "stage_name": "permutation_cycle", "status": "training", "step": 193, "tokens_per_sec_per_gpu": 99012.570796, "vram_gb": 20.709735} {"adam_lr": 9.75e-05, "dt_ms": 162.538528, "global_tokens_per_sec": 100800.715726, "grad_norm": 46.25, "loss": 2.844495, "lr_stage": "warmup", "mfu": 57.903816, "stage_name": "permutation_cycle", "status": "training", "step": 194, "tokens_per_sec_per_gpu": 100800.715726, "vram_gb": 20.709735} {"adam_lr": 9.8e-05, "dt_ms": 161.985159, "global_tokens_per_sec": 101145.068531, "grad_norm": 49.5, "loss": 3.056978, "lr_stage": "warmup", "mfu": 58.101626, "stage_name": "permutation_cycle", "status": "training", "step": 195, "tokens_per_sec_per_gpu": 101145.068531, "vram_gb": 20.709735} {"adam_lr": 9.850000000000001e-05, "dt_ms": 162.726641, "global_tokens_per_sec": 100684.189936, "grad_norm": 40.25, "loss": 2.521211, "lr_stage": "warmup", "mfu": 57.836879, "stage_name": "permutation_cycle", "status": "training", "step": 196, "tokens_per_sec_per_gpu": 100684.189936, "vram_gb": 20.709735} {"adam_lr": 9.900000000000001e-05, "dt_ms": 164.702177, "global_tokens_per_sec": 99476.523587, "grad_norm": 44.0, "loss": 3.406513, "lr_stage": "warmup", "mfu": 57.143149, "stage_name": "permutation_cycle", "status": "training", "step": 197, "tokens_per_sec_per_gpu": 99476.523587, "vram_gb": 20.709735} {"adam_lr": 9.95e-05, "dt_ms": 162.642479, "global_tokens_per_sec": 100736.29046, "grad_norm": 46.5, "loss": 3.384606, "lr_stage": "warmup", "mfu": 57.866808, "stage_name": "permutation_cycle", "status": "training", "step": 198, "tokens_per_sec_per_gpu": 100736.29046, "vram_gb": 20.709735} {"adam_lr": 0.0001, "dt_ms": 164.437056, "global_tokens_per_sec": 99636.909342, "grad_norm": 40.0, "loss": 3.196893, "lr_stage": "warmup", "mfu": 57.235281, "stage_name": "permutation_cycle", "status": "training", "step": 199, "tokens_per_sec_per_gpu": 99636.909342, "vram_gb": 20.709735} {"adam_lr": 0.0001, "dt_ms": 162.942886, "global_tokens_per_sec": 100550.569385, "grad_norm": 52.0, "loss": 3.471493, "lr_stage": "cosine_decay", "mfu": 57.760123, "stage_name": "permutation_cycle", "status": "training", "step": 200, "tokens_per_sec_per_gpu": 100550.569385, "vram_gb": 20.709735} {"adam_lr": 9.999992384566439e-05, "dt_ms": 156.53038, "global_tokens_per_sec": 104669.777036, "grad_norm": 46.25, "loss": 3.425111, "lr_stage": "cosine_decay", "mfu": 60.126354, "stage_name": "permutation_cycle", "status": "training", "step": 201, "tokens_per_sec_per_gpu": 104669.777036, "vram_gb": 20.709735} {"adam_lr": 9.999969538288952e-05, "dt_ms": 159.310818, "global_tokens_per_sec": 102842.984768, "grad_norm": 54.25, "loss": 3.592, "lr_stage": "cosine_decay", "mfu": 59.076974, "stage_name": "permutation_cycle", "status": "training", "step": 202, "tokens_per_sec_per_gpu": 102842.984768, "vram_gb": 20.709735} {"adam_lr": 9.999931461237134e-05, "dt_ms": 159.25765, "global_tokens_per_sec": 102877.318367, "grad_norm": 51.75, "loss": 3.081065, "lr_stage": "cosine_decay", "mfu": 59.096697, "stage_name": "permutation_cycle", "status": "training", "step": 203, "tokens_per_sec_per_gpu": 102877.318367, "vram_gb": 20.709735} {"adam_lr": 9.999878153526974e-05, "dt_ms": 159.63316, "global_tokens_per_sec": 102635.317356, "grad_norm": 48.0, "loss": 3.769625, "lr_stage": "cosine_decay", "mfu": 58.957682, "stage_name": "permutation_cycle", "status": "training", "step": 204, "tokens_per_sec_per_gpu": 102635.317356, "vram_gb": 20.709735} {"adam_lr": 9.999809615320856e-05, "dt_ms": 158.583164, "global_tokens_per_sec": 103314.876337, "grad_norm": 35.0, "loss": 1.84017, "lr_stage": "cosine_decay", "mfu": 59.348047, "stage_name": "permutation_cycle", "status": "training", "step": 205, "tokens_per_sec_per_gpu": 103314.876337, "vram_gb": 20.709735} {"adam_lr": 9.999725846827562e-05, "dt_ms": 175.095081, "global_tokens_per_sec": 93572.017418, "grad_norm": 42.75, "loss": 2.839034, "lr_stage": "cosine_decay", "mfu": 53.751373, "stage_name": "permutation_cycle", "status": "training", "step": 206, "tokens_per_sec_per_gpu": 93572.017418, "vram_gb": 20.709735} {"adam_lr": 9.999626848302261e-05, "dt_ms": 164.408684, "global_tokens_per_sec": 99654.103564, "grad_norm": 56.25, "loss": 3.500397, "lr_stage": "cosine_decay", "mfu": 57.245158, "stage_name": "permutation_cycle", "status": "training", "step": 207, "tokens_per_sec_per_gpu": 99654.103564, "vram_gb": 20.709735} {"adam_lr": 9.999512620046522e-05, "dt_ms": 161.635876, "global_tokens_per_sec": 101363.635572, "grad_norm": 44.5, "loss": 3.40796, "lr_stage": "cosine_decay", "mfu": 58.227179, "stage_name": "permutation_cycle", "status": "training", "step": 208, "tokens_per_sec_per_gpu": 101363.635572, "vram_gb": 20.709735} {"adam_lr": 9.999383162408304e-05, "dt_ms": 159.130573, "global_tokens_per_sec": 102959.473237, "grad_norm": 40.0, "loss": 3.722782, "lr_stage": "cosine_decay", "mfu": 59.14389, "stage_name": "permutation_cycle", "status": "training", "step": 209, "tokens_per_sec_per_gpu": 102959.473237, "vram_gb": 20.709735} {"adam_lr": 9.999238475781957e-05, "dt_ms": 160.843372, "global_tokens_per_sec": 101863.071889, "grad_norm": 45.5, "loss": 3.920818, "lr_stage": "cosine_decay", "mfu": 58.514075, "stage_name": "permutation_cycle", "status": "training", "step": 210, "tokens_per_sec_per_gpu": 101863.071889, "vram_gb": 20.709735} {"adam_lr": 9.999078560608222e-05, "dt_ms": 159.386158, "global_tokens_per_sec": 102794.371899, "grad_norm": 44.5, "loss": 2.827234, "lr_stage": "cosine_decay", "mfu": 59.049049, "stage_name": "permutation_cycle", "status": "training", "step": 211, "tokens_per_sec_per_gpu": 102794.371899, "vram_gb": 20.709735} {"adam_lr": 9.998903417374228e-05, "dt_ms": 170.958281, "global_tokens_per_sec": 95836.246984, "grad_norm": 50.75, "loss": 2.479997, "lr_stage": "cosine_decay", "mfu": 55.052034, "stage_name": "permutation_cycle", "status": "training", "step": 212, "tokens_per_sec_per_gpu": 95836.246984, "vram_gb": 20.709735} {"adam_lr": 9.998713046613493e-05, "dt_ms": 164.749384, "global_tokens_per_sec": 99448.019832, "grad_norm": 40.5, "loss": 2.196865, "lr_stage": "cosine_decay", "mfu": 57.126776, "stage_name": "permutation_cycle", "status": "training", "step": 213, "tokens_per_sec_per_gpu": 99448.019832, "vram_gb": 20.709735} {"adam_lr": 9.998507448905917e-05, "dt_ms": 163.883209, "global_tokens_per_sec": 99973.634133, "grad_norm": 44.5, "loss": 3.909203, "lr_stage": "cosine_decay", "mfu": 57.428709, "stage_name": "permutation_cycle", "status": "training", "step": 214, "tokens_per_sec_per_gpu": 99973.634133, "vram_gb": 20.709735} {"adam_lr": 9.998286624877786e-05, "dt_ms": 159.059286, "global_tokens_per_sec": 103005.61759, "grad_norm": 54.5, "loss": 2.977192, "lr_stage": "cosine_decay", "mfu": 59.170397, "stage_name": "permutation_cycle", "status": "training", "step": 215, "tokens_per_sec_per_gpu": 103005.61759, "vram_gb": 20.709735} {"adam_lr": 9.998050575201771e-05, "dt_ms": 158.499718, "global_tokens_per_sec": 103369.269274, "grad_norm": 52.75, "loss": 3.421298, "lr_stage": "cosine_decay", "mfu": 59.379292, "stage_name": "permutation_cycle", "status": "training", "step": 216, "tokens_per_sec_per_gpu": 103369.269274, "vram_gb": 20.709735} {"adam_lr": 9.997799300596921e-05, "dt_ms": 161.896467, "global_tokens_per_sec": 101200.478815, "grad_norm": 59.5, "loss": 3.685323, "lr_stage": "cosine_decay", "mfu": 58.133456, "stage_name": "permutation_cycle", "status": "training", "step": 217, "tokens_per_sec_per_gpu": 101200.478815, "vram_gb": 20.709735} {"adam_lr": 9.997532801828658e-05, "dt_ms": 165.463686, "global_tokens_per_sec": 99018.705537, "grad_norm": 72.5, "loss": 2.71441, "lr_stage": "cosine_decay", "mfu": 56.880161, "stage_name": "permutation_cycle", "status": "training", "step": 218, "tokens_per_sec_per_gpu": 99018.705537, "vram_gb": 20.709735} {"adam_lr": 9.997251079708787e-05, "dt_ms": 166.583538, "global_tokens_per_sec": 98353.055718, "grad_norm": 38.25, "loss": 3.189452, "lr_stage": "cosine_decay", "mfu": 56.497786, "stage_name": "permutation_cycle", "status": "training", "step": 219, "tokens_per_sec_per_gpu": 98353.055718, "vram_gb": 20.709735} {"adam_lr": 9.99695413509548e-05, "dt_ms": 163.32984, "global_tokens_per_sec": 100312.349718, "grad_norm": 52.75, "loss": 3.619444, "lr_stage": "cosine_decay", "mfu": 57.62328, "stage_name": "permutation_cycle", "status": "training", "step": 220, "tokens_per_sec_per_gpu": 100312.349718, "vram_gb": 20.709735} {"adam_lr": 9.996641968893282e-05, "dt_ms": 162.975073, "global_tokens_per_sec": 100530.711307, "grad_norm": 41.75, "loss": 3.408341, "lr_stage": "cosine_decay", "mfu": 57.748715, "stage_name": "permutation_cycle", "status": "training", "step": 221, "tokens_per_sec_per_gpu": 100530.711307, "vram_gb": 20.709735} {"adam_lr": 9.996314582053106e-05, "dt_ms": 160.479307, "global_tokens_per_sec": 102094.159605, "grad_norm": 44.0, "loss": 3.378567, "lr_stage": "cosine_decay", "mfu": 58.64682, "stage_name": "permutation_cycle", "status": "training", "step": 222, "tokens_per_sec_per_gpu": 102094.159605, "vram_gb": 20.709735} {"adam_lr": 9.99597197557223e-05, "dt_ms": 162.513256, "global_tokens_per_sec": 100816.391203, "grad_norm": 44.75, "loss": 4.065776, "lr_stage": "cosine_decay", "mfu": 57.912821, "stage_name": "permutation_cycle", "status": "training", "step": 223, "tokens_per_sec_per_gpu": 100816.391203, "vram_gb": 20.709735} {"adam_lr": 9.995614150494293e-05, "dt_ms": 164.837837, "global_tokens_per_sec": 99394.655234, "grad_norm": 39.0, "loss": 2.787323, "lr_stage": "cosine_decay", "mfu": 57.096121, "stage_name": "permutation_cycle", "status": "training", "step": 224, "tokens_per_sec_per_gpu": 99394.655234, "vram_gb": 20.709735} {"adam_lr": 9.99524110790929e-05, "dt_ms": 165.457487, "global_tokens_per_sec": 99022.415283, "grad_norm": 42.5, "loss": 3.181653, "lr_stage": "cosine_decay", "mfu": 56.882292, "stage_name": "permutation_cycle", "status": "training", "step": 225, "tokens_per_sec_per_gpu": 99022.415283, "vram_gb": 20.709735} {"adam_lr": 9.994852848953574e-05, "dt_ms": 161.767483, "global_tokens_per_sec": 101281.170484, "grad_norm": 38.25, "loss": 3.443936, "lr_stage": "cosine_decay", "mfu": 58.179808, "stage_name": "permutation_cycle", "status": "training", "step": 226, "tokens_per_sec_per_gpu": 101281.170484, "vram_gb": 20.709735} {"adam_lr": 9.994449374809851e-05, "dt_ms": 162.649632, "global_tokens_per_sec": 100731.860557, "grad_norm": 50.0, "loss": 3.136158, "lr_stage": "cosine_decay", "mfu": 57.864263, "stage_name": "permutation_cycle", "status": "training", "step": 227, "tokens_per_sec_per_gpu": 100731.860557, "vram_gb": 20.709735} {"adam_lr": 9.99403068670717e-05, "dt_ms": 161.00502, "global_tokens_per_sec": 101760.802151, "grad_norm": 53.5, "loss": 3.197154, "lr_stage": "cosine_decay", "mfu": 58.455327, "stage_name": "permutation_cycle", "status": "training", "step": 228, "tokens_per_sec_per_gpu": 101760.802151, "vram_gb": 20.709735} {"adam_lr": 9.99359678592093e-05, "dt_ms": 162.673235, "global_tokens_per_sec": 100717.244641, "grad_norm": 87.5, "loss": 4.807237, "lr_stage": "cosine_decay", "mfu": 57.855867, "stage_name": "permutation_cycle", "status": "training", "step": 229, "tokens_per_sec_per_gpu": 100717.244641, "vram_gb": 20.709735} {"adam_lr": 9.99314767377287e-05, "dt_ms": 164.947271, "global_tokens_per_sec": 99328.711934, "grad_norm": 53.5, "loss": 3.033283, "lr_stage": "cosine_decay", "mfu": 57.05824, "stage_name": "permutation_cycle", "status": "training", "step": 230, "tokens_per_sec_per_gpu": 99328.711934, "vram_gb": 20.709735} {"adam_lr": 9.99268335163106e-05, "dt_ms": 163.364172, "global_tokens_per_sec": 100291.268283, "grad_norm": 42.75, "loss": 3.307084, "lr_stage": "cosine_decay", "mfu": 57.61117, "stage_name": "permutation_cycle", "status": "training", "step": 231, "tokens_per_sec_per_gpu": 100291.268283, "vram_gb": 20.709735} {"adam_lr": 9.992203820909906e-05, "dt_ms": 162.51421, "global_tokens_per_sec": 100815.799587, "grad_norm": 44.25, "loss": 4.185788, "lr_stage": "cosine_decay", "mfu": 57.912481, "stage_name": "permutation_cycle", "status": "training", "step": 232, "tokens_per_sec_per_gpu": 100815.799587, "vram_gb": 20.709735} {"adam_lr": 9.991709083070143e-05, "dt_ms": 160.610914, "global_tokens_per_sec": 102010.502079, "grad_norm": 53.5, "loss": 4.767833, "lr_stage": "cosine_decay", "mfu": 58.598764, "stage_name": "permutation_cycle", "status": "training", "step": 233, "tokens_per_sec_per_gpu": 102010.502079, "vram_gb": 20.709735} {"adam_lr": 9.991199139618827e-05, "dt_ms": 161.808968, "global_tokens_per_sec": 101255.203862, "grad_norm": 53.0, "loss": 3.916497, "lr_stage": "cosine_decay", "mfu": 58.164892, "stage_name": "permutation_cycle", "status": "training", "step": 234, "tokens_per_sec_per_gpu": 101255.203862, "vram_gb": 20.709735} {"adam_lr": 9.990673992109335e-05, "dt_ms": 161.530495, "global_tokens_per_sec": 101429.764277, "grad_norm": 66.0, "loss": 3.755562, "lr_stage": "cosine_decay", "mfu": 58.265166, "stage_name": "permutation_cycle", "status": "training", "step": 235, "tokens_per_sec_per_gpu": 101429.764277, "vram_gb": 20.709735} {"adam_lr": 9.990133642141359e-05, "dt_ms": 166.46862, "global_tokens_per_sec": 98420.951471, "grad_norm": 38.25, "loss": 3.467594, "lr_stage": "cosine_decay", "mfu": 56.536788, "stage_name": "permutation_cycle", "status": "training", "step": 236, "tokens_per_sec_per_gpu": 98420.951471, "vram_gb": 20.709735} {"adam_lr": 9.989578091360896e-05, "dt_ms": 164.672852, "global_tokens_per_sec": 99494.238695, "grad_norm": 53.75, "loss": 3.233621, "lr_stage": "cosine_decay", "mfu": 57.153325, "stage_name": "permutation_cycle", "status": "training", "step": 237, "tokens_per_sec_per_gpu": 99494.238695, "vram_gb": 20.709735} {"adam_lr": 9.98900734146025e-05, "dt_ms": 162.554502, "global_tokens_per_sec": 100790.810155, "grad_norm": 48.25, "loss": 3.46314, "lr_stage": "cosine_decay", "mfu": 57.898126, "stage_name": "permutation_cycle", "status": "training", "step": 238, "tokens_per_sec_per_gpu": 100790.810155, "vram_gb": 20.709735} {"adam_lr": 9.988421394178027e-05, "dt_ms": 160.694599, "global_tokens_per_sec": 101957.378073, "grad_norm": 61.75, "loss": 2.822249, "lr_stage": "cosine_decay", "mfu": 58.568248, "stage_name": "permutation_cycle", "status": "training", "step": 239, "tokens_per_sec_per_gpu": 101957.378073, "vram_gb": 20.709735} {"adam_lr": 9.987820251299122e-05, "dt_ms": 162.979126, "global_tokens_per_sec": 100528.211216, "grad_norm": 44.0, "loss": 3.743526, "lr_stage": "cosine_decay", "mfu": 57.747279, "stage_name": "permutation_cycle", "status": "training", "step": 240, "tokens_per_sec_per_gpu": 100528.211216, "vram_gb": 20.709735} {"adam_lr": 9.98720391465472e-05, "dt_ms": 160.978079, "global_tokens_per_sec": 101777.832844, "grad_norm": 72.0, "loss": 3.347439, "lr_stage": "cosine_decay", "mfu": 58.46511, "stage_name": "permutation_cycle", "status": "training", "step": 241, "tokens_per_sec_per_gpu": 101777.832844, "vram_gb": 20.709735} {"adam_lr": 9.986572386122291e-05, "dt_ms": 164.009809, "global_tokens_per_sec": 99896.463819, "grad_norm": 45.25, "loss": 3.98051, "lr_stage": "cosine_decay", "mfu": 57.384379, "stage_name": "permutation_cycle", "status": "training", "step": 242, "tokens_per_sec_per_gpu": 99896.463819, "vram_gb": 20.709735} {"adam_lr": 9.985925667625579e-05, "dt_ms": 165.707588, "global_tokens_per_sec": 98872.961573, "grad_norm": 64.5, "loss": 3.393056, "lr_stage": "cosine_decay", "mfu": 56.79644, "stage_name": "permutation_cycle", "status": "training", "step": 243, "tokens_per_sec_per_gpu": 98872.961573, "vram_gb": 20.709735} {"adam_lr": 9.985263761134602e-05, "dt_ms": 161.996126, "global_tokens_per_sec": 101138.220937, "grad_norm": 62.0, "loss": 4.243887, "lr_stage": "cosine_decay", "mfu": 58.097692, "stage_name": "permutation_cycle", "status": "training", "step": 244, "tokens_per_sec_per_gpu": 101138.220937, "vram_gb": 20.709735} {"adam_lr": 9.98458666866564e-05, "dt_ms": 161.220312, "global_tokens_per_sec": 101624.911804, "grad_norm": 40.0, "loss": 3.141117, "lr_stage": "cosine_decay", "mfu": 58.377266, "stage_name": "permutation_cycle", "status": "training", "step": 245, "tokens_per_sec_per_gpu": 101624.911804, "vram_gb": 20.709735} {"adam_lr": 9.983894392281237e-05, "dt_ms": 162.662745, "global_tokens_per_sec": 100723.74008, "grad_norm": 42.25, "loss": 3.838413, "lr_stage": "cosine_decay", "mfu": 57.859598, "stage_name": "permutation_cycle", "status": "training", "step": 246, "tokens_per_sec_per_gpu": 100723.74008, "vram_gb": 20.709735} {"adam_lr": 9.983186934090182e-05, "dt_ms": 163.063288, "global_tokens_per_sec": 100476.32565, "grad_norm": 47.0, "loss": 3.245317, "lr_stage": "cosine_decay", "mfu": 57.717474, "stage_name": "permutation_cycle", "status": "training", "step": 247, "tokens_per_sec_per_gpu": 100476.32565, "vram_gb": 20.709735} {"adam_lr": 9.982464296247522e-05, "dt_ms": 162.560225, "global_tokens_per_sec": 100787.262364, "grad_norm": 49.0, "loss": 4.098408, "lr_stage": "cosine_decay", "mfu": 57.896088, "stage_name": "permutation_cycle", "status": "training", "step": 248, "tokens_per_sec_per_gpu": 100787.262364, "vram_gb": 20.709735} {"adam_lr": 9.981726480954531e-05, "dt_ms": 162.894011, "global_tokens_per_sec": 100580.739251, "grad_norm": 35.25, "loss": 2.215635, "lr_stage": "cosine_decay", "mfu": 57.777453, "stage_name": "permutation_cycle", "status": "training", "step": 249, "tokens_per_sec_per_gpu": 100580.739251, "vram_gb": 20.709735} {"adam_lr": 9.980973490458728e-05, "dt_ms": 163.975239, "global_tokens_per_sec": 99917.524865, "grad_norm": 45.25, "loss": 3.673236, "lr_stage": "cosine_decay", "mfu": 57.396477, "stage_name": "permutation_cycle", "status": "training", "step": 250, "tokens_per_sec_per_gpu": 99917.524865, "vram_gb": 20.709735} {"adam_lr": 9.980205327053848e-05, "dt_ms": 163.785458, "global_tokens_per_sec": 100033.301118, "grad_norm": 39.5, "loss": 3.162615, "lr_stage": "cosine_decay", "mfu": 57.462984, "stage_name": "permutation_cycle", "status": "training", "step": 251, "tokens_per_sec_per_gpu": 100033.301118, "vram_gb": 20.709735} {"adam_lr": 9.979421993079852e-05, "dt_ms": 161.199093, "global_tokens_per_sec": 101638.28908, "grad_norm": 56.0, "loss": 3.257223, "lr_stage": "cosine_decay", "mfu": 58.384951, "stage_name": "permutation_cycle", "status": "training", "step": 252, "tokens_per_sec_per_gpu": 101638.28908, "vram_gb": 20.709735} {"adam_lr": 9.978623490922911e-05, "dt_ms": 164.797306, "global_tokens_per_sec": 99419.100904, "grad_norm": 53.25, "loss": 3.629247, "lr_stage": "cosine_decay", "mfu": 57.110163, "stage_name": "permutation_cycle", "status": "training", "step": 253, "tokens_per_sec_per_gpu": 99419.100904, "vram_gb": 20.709735} {"adam_lr": 9.977809823015401e-05, "dt_ms": 162.271023, "global_tokens_per_sec": 100966.886864, "grad_norm": 39.75, "loss": 2.917455, "lr_stage": "cosine_decay", "mfu": 57.999271, "stage_name": "permutation_cycle", "status": "training", "step": 254, "tokens_per_sec_per_gpu": 100966.886864, "vram_gb": 20.709735} {"adam_lr": 9.976980991835894e-05, "dt_ms": 162.978888, "global_tokens_per_sec": 100528.358277, "grad_norm": 43.5, "loss": 2.388982, "lr_stage": "cosine_decay", "mfu": 57.747364, "stage_name": "permutation_cycle", "status": "training", "step": 255, "tokens_per_sec_per_gpu": 100528.358277, "vram_gb": 20.709735} {"adam_lr": 9.976136999909156e-05, "dt_ms": 390.392065, "global_tokens_per_sec": 41968.066123, "grad_norm": 58.0, "loss": 4.000111, "lr_stage": "cosine_decay", "mfu": 24.108075, "stage_name": "permutation_cycle", "status": "training", "step": 256, "tokens_per_sec_per_gpu": 41968.066123, "vram_gb": 20.709735} {"adam_lr": 9.975277849806133e-05, "dt_ms": 159.35111, "global_tokens_per_sec": 102816.980396, "grad_norm": 46.5, "loss": 3.814317, "lr_stage": "cosine_decay", "mfu": 59.062036, "stage_name": "permutation_cycle", "status": "training", "step": 257, "tokens_per_sec_per_gpu": 102816.980396, "vram_gb": 20.709735} {"adam_lr": 9.974403544143941e-05, "dt_ms": 159.386158, "global_tokens_per_sec": 102794.371899, "grad_norm": 48.5, "loss": 4.029976, "lr_stage": "cosine_decay", "mfu": 59.049049, "stage_name": "permutation_cycle", "status": "training", "step": 258, "tokens_per_sec_per_gpu": 102794.371899, "vram_gb": 20.709735} {"adam_lr": 9.973514085585871e-05, "dt_ms": 158.744812, "global_tokens_per_sec": 103209.67213, "grad_norm": 61.5, "loss": 4.175882, "lr_stage": "cosine_decay", "mfu": 59.287614, "stage_name": "permutation_cycle", "status": "training", "step": 259, "tokens_per_sec_per_gpu": 103209.67213, "vram_gb": 20.709735} {"adam_lr": 9.972609476841367e-05, "dt_ms": 158.322573, "global_tokens_per_sec": 103484.927763, "grad_norm": 38.25, "loss": 2.60753, "lr_stage": "cosine_decay", "mfu": 59.445731, "stage_name": "permutation_cycle", "status": "training", "step": 260, "tokens_per_sec_per_gpu": 103484.927763, "vram_gb": 20.709735} {"adam_lr": 9.971689720666024e-05, "dt_ms": 175.38476, "global_tokens_per_sec": 93417.466883, "grad_norm": 51.25, "loss": 3.303433, "lr_stage": "cosine_decay", "mfu": 53.662594, "stage_name": "permutation_cycle", "status": "training", "step": 261, "tokens_per_sec_per_gpu": 93417.466883, "vram_gb": 20.709735} {"adam_lr": 9.970754819861577e-05, "dt_ms": 167.413473, "global_tokens_per_sec": 97865.48056, "grad_norm": 64.5, "loss": 4.816503, "lr_stage": "cosine_decay", "mfu": 56.217704, "stage_name": "permutation_cycle", "status": "training", "step": 262, "tokens_per_sec_per_gpu": 97865.48056, "vram_gb": 20.709735} {"adam_lr": 9.9698047772759e-05, "dt_ms": 159.75976, "global_tokens_per_sec": 102553.98487, "grad_norm": 42.25, "loss": 2.787268, "lr_stage": "cosine_decay", "mfu": 58.910962, "stage_name": "permutation_cycle", "status": "training", "step": 263, "tokens_per_sec_per_gpu": 102553.98487, "vram_gb": 20.709735} {"adam_lr": 9.968839595802982e-05, "dt_ms": 161.385775, "global_tokens_per_sec": 101520.719774, "grad_norm": 54.0, "loss": 4.409842, "lr_stage": "cosine_decay", "mfu": 58.317414, "stage_name": "permutation_cycle", "status": "training", "step": 264, "tokens_per_sec_per_gpu": 101520.719774, "vram_gb": 20.709735} {"adam_lr": 9.967859278382938e-05, "dt_ms": 158.803225, "global_tokens_per_sec": 103171.708541, "grad_norm": 37.75, "loss": 2.886565, "lr_stage": "cosine_decay", "mfu": 59.265806, "stage_name": "permutation_cycle", "status": "training", "step": 265, "tokens_per_sec_per_gpu": 103171.708541, "vram_gb": 20.709735} {"adam_lr": 9.966863828001982e-05, "dt_ms": 160.432339, "global_tokens_per_sec": 102124.048875, "grad_norm": 58.5, "loss": 3.478157, "lr_stage": "cosine_decay", "mfu": 58.66399, "stage_name": "permutation_cycle", "status": "training", "step": 266, "tokens_per_sec_per_gpu": 102124.048875, "vram_gb": 20.709735} {"adam_lr": 9.965853247692432e-05, "dt_ms": 172.395945, "global_tokens_per_sec": 95037.038362, "grad_norm": 46.0, "loss": 4.167708, "lr_stage": "cosine_decay", "mfu": 54.592938, "stage_name": "permutation_cycle", "status": "training", "step": 267, "tokens_per_sec_per_gpu": 95037.038362, "vram_gb": 20.709735} {"adam_lr": 9.964827540532685e-05, "dt_ms": 165.272713, "global_tokens_per_sec": 99133.122048, "grad_norm": 46.0, "loss": 4.144051, "lr_stage": "cosine_decay", "mfu": 56.945886, "stage_name": "permutation_cycle", "status": "training", "step": 268, "tokens_per_sec_per_gpu": 99133.122048, "vram_gb": 20.709735} {"adam_lr": 9.963786709647228e-05, "dt_ms": 162.424564, "global_tokens_per_sec": 100871.441856, "grad_norm": 41.75, "loss": 3.488543, "lr_stage": "cosine_decay", "mfu": 57.944444, "stage_name": "permutation_cycle", "status": "training", "step": 269, "tokens_per_sec_per_gpu": 100871.441856, "vram_gb": 20.709735} {"adam_lr": 9.962730758206611e-05, "dt_ms": 161.956549, "global_tokens_per_sec": 101162.936185, "grad_norm": 46.5, "loss": 3.627921, "lr_stage": "cosine_decay", "mfu": 58.11189, "stage_name": "permutation_cycle", "status": "training", "step": 270, "tokens_per_sec_per_gpu": 101162.936185, "vram_gb": 20.709735} {"adam_lr": 9.961659689427443e-05, "dt_ms": 159.808874, "global_tokens_per_sec": 102522.466848, "grad_norm": 30.625, "loss": 2.809833, "lr_stage": "cosine_decay", "mfu": 58.892856, "stage_name": "permutation_cycle", "status": "training", "step": 271, "tokens_per_sec_per_gpu": 102522.466848, "vram_gb": 20.709735} {"adam_lr": 9.96057350657239e-05, "dt_ms": 162.372351, "global_tokens_per_sec": 100903.878832, "grad_norm": 37.75, "loss": 3.80465, "lr_stage": "cosine_decay", "mfu": 57.963077, "stage_name": "permutation_cycle", "status": "training", "step": 272, "tokens_per_sec_per_gpu": 100903.878832, "vram_gb": 20.709735} {"adam_lr": 9.95947221295015e-05, "dt_ms": 167.940617, "global_tokens_per_sec": 97558.293705, "grad_norm": 66.0, "loss": 4.033961, "lr_stage": "cosine_decay", "mfu": 56.041244, "stage_name": "permutation_cycle", "status": "training", "step": 273, "tokens_per_sec_per_gpu": 97558.293705, "vram_gb": 20.709735} {"adam_lr": 9.958355811915451e-05, "dt_ms": 163.869858, "global_tokens_per_sec": 99981.779573, "grad_norm": 46.0, "loss": 3.745104, "lr_stage": "cosine_decay", "mfu": 57.433388, "stage_name": "permutation_cycle", "status": "training", "step": 274, "tokens_per_sec_per_gpu": 99981.779573, "vram_gb": 20.709735} {"adam_lr": 9.957224306869053e-05, "dt_ms": 161.220074, "global_tokens_per_sec": 101625.062091, "grad_norm": 51.5, "loss": 3.461848, "lr_stage": "cosine_decay", "mfu": 58.377353, "stage_name": "permutation_cycle", "status": "training", "step": 275, "tokens_per_sec_per_gpu": 101625.062091, "vram_gb": 20.709735} {"adam_lr": 9.956077701257709e-05, "dt_ms": 159.4944, "global_tokens_per_sec": 102724.609751, "grad_norm": 49.25, "loss": 2.765469, "lr_stage": "cosine_decay", "mfu": 59.008975, "stage_name": "permutation_cycle", "status": "training", "step": 276, "tokens_per_sec_per_gpu": 102724.609751, "vram_gb": 20.709735} {"adam_lr": 9.954915998574181e-05, "dt_ms": 160.132647, "global_tokens_per_sec": 102315.176523, "grad_norm": 55.0, "loss": 2.956883, "lr_stage": "cosine_decay", "mfu": 58.773781, "stage_name": "permutation_cycle", "status": "training", "step": 277, "tokens_per_sec_per_gpu": 102315.176523, "vram_gb": 20.709735} {"adam_lr": 9.953739202357218e-05, "dt_ms": 164.900064, "global_tokens_per_sec": 99357.147329, "grad_norm": 38.5, "loss": 3.618833, "lr_stage": "cosine_decay", "mfu": 57.074575, "stage_name": "permutation_cycle", "status": "training", "step": 278, "tokens_per_sec_per_gpu": 99357.147329, "vram_gb": 20.709735} {"adam_lr": 9.952547316191544e-05, "dt_ms": 165.492773, "global_tokens_per_sec": 99001.301975, "grad_norm": 41.5, "loss": 3.548382, "lr_stage": "cosine_decay", "mfu": 56.870164, "stage_name": "permutation_cycle", "status": "training", "step": 279, "tokens_per_sec_per_gpu": 99001.301975, "vram_gb": 20.709735} {"adam_lr": 9.951340343707852e-05, "dt_ms": 165.616751, "global_tokens_per_sec": 98927.191417, "grad_norm": 46.75, "loss": 3.483757, "lr_stage": "cosine_decay", "mfu": 56.827592, "stage_name": "permutation_cycle", "status": "training", "step": 280, "tokens_per_sec_per_gpu": 98927.191417, "vram_gb": 20.709735} {"adam_lr": 9.950118288582788e-05, "dt_ms": 162.881851, "global_tokens_per_sec": 100588.247737, "grad_norm": 38.0, "loss": 4.10147, "lr_stage": "cosine_decay", "mfu": 57.781766, "stage_name": "permutation_cycle", "status": "training", "step": 281, "tokens_per_sec_per_gpu": 100588.247737, "vram_gb": 20.709735} {"adam_lr": 9.948881154538945e-05, "dt_ms": 161.564589, "global_tokens_per_sec": 101408.360256, "grad_norm": 49.0, "loss": 3.761658, "lr_stage": "cosine_decay", "mfu": 58.252871, "stage_name": "permutation_cycle", "status": "training", "step": 282, "tokens_per_sec_per_gpu": 101408.360256, "vram_gb": 20.709735} {"adam_lr": 9.947628945344848e-05, "dt_ms": 161.61561, "global_tokens_per_sec": 101376.345933, "grad_norm": 46.75, "loss": 3.845845, "lr_stage": "cosine_decay", "mfu": 58.23448, "stage_name": "permutation_cycle", "status": "training", "step": 283, "tokens_per_sec_per_gpu": 101376.345933, "vram_gb": 20.709735} {"adam_lr": 9.946361664814943e-05, "dt_ms": 163.988829, "global_tokens_per_sec": 99909.244636, "grad_norm": 29.875, "loss": 2.8139, "lr_stage": "cosine_decay", "mfu": 57.391721, "stage_name": "permutation_cycle", "status": "training", "step": 284, "tokens_per_sec_per_gpu": 99909.244636, "vram_gb": 20.709735} {"adam_lr": 9.945079316809585e-05, "dt_ms": 165.351868, "global_tokens_per_sec": 99085.666405, "grad_norm": 34.5, "loss": 2.712401, "lr_stage": "cosine_decay", "mfu": 56.918626, "stage_name": "permutation_cycle", "status": "training", "step": 285, "tokens_per_sec_per_gpu": 99085.666405, "vram_gb": 20.709735} {"adam_lr": 9.94378190523503e-05, "dt_ms": 164.188385, "global_tokens_per_sec": 99787.813852, "grad_norm": 43.5, "loss": 2.845379, "lr_stage": "cosine_decay", "mfu": 57.321966, "stage_name": "permutation_cycle", "status": "training", "step": 286, "tokens_per_sec_per_gpu": 99787.813852, "vram_gb": 20.709735} {"adam_lr": 9.942469434043418e-05, "dt_ms": 162.003279, "global_tokens_per_sec": 101133.755614, "grad_norm": 49.25, "loss": 3.041309, "lr_stage": "cosine_decay", "mfu": 58.095127, "stage_name": "permutation_cycle", "status": "training", "step": 287, "tokens_per_sec_per_gpu": 101133.755614, "vram_gb": 20.709735} {"adam_lr": 9.941141907232765e-05, "dt_ms": 161.481142, "global_tokens_per_sec": 101460.763731, "grad_norm": 48.75, "loss": 3.338633, "lr_stage": "cosine_decay", "mfu": 58.282973, "stage_name": "permutation_cycle", "status": "training", "step": 288, "tokens_per_sec_per_gpu": 101460.763731, "vram_gb": 20.709735} {"adam_lr": 9.939799328846946e-05, "dt_ms": 162.423849, "global_tokens_per_sec": 100871.886057, "grad_norm": 66.5, "loss": 4.086764, "lr_stage": "cosine_decay", "mfu": 57.944699, "stage_name": "permutation_cycle", "status": "training", "step": 289, "tokens_per_sec_per_gpu": 100871.886057, "vram_gb": 20.709735} {"adam_lr": 9.938441702975689e-05, "dt_ms": 161.337852, "global_tokens_per_sec": 101550.874444, "grad_norm": 48.5, "loss": 3.368001, "lr_stage": "cosine_decay", "mfu": 58.334736, "stage_name": "permutation_cycle", "status": "training", "step": 290, "tokens_per_sec_per_gpu": 101550.874444, "vram_gb": 20.709735} {"adam_lr": 9.937069033754558e-05, "dt_ms": 163.743019, "global_tokens_per_sec": 100059.2275, "grad_norm": 69.5, "loss": 4.194294, "lr_stage": "cosine_decay", "mfu": 57.477877, "stage_name": "permutation_cycle", "status": "training", "step": 291, "tokens_per_sec_per_gpu": 100059.2275, "vram_gb": 20.709735} {"adam_lr": 9.93568132536494e-05, "dt_ms": 163.145781, "global_tokens_per_sec": 100425.520926, "grad_norm": 50.0, "loss": 4.17883, "lr_stage": "cosine_decay", "mfu": 57.68829, "stage_name": "permutation_cycle", "status": "training", "step": 292, "tokens_per_sec_per_gpu": 100425.520926, "vram_gb": 20.709735} {"adam_lr": 9.934278582034037e-05, "dt_ms": 163.622379, "global_tokens_per_sec": 100133.001792, "grad_norm": 41.25, "loss": 3.932523, "lr_stage": "cosine_decay", "mfu": 57.520256, "stage_name": "permutation_cycle", "status": "training", "step": 293, "tokens_per_sec_per_gpu": 100133.001792, "vram_gb": 20.709735} {"adam_lr": 9.932860808034848e-05, "dt_ms": 161.793709, "global_tokens_per_sec": 101264.753255, "grad_norm": 38.75, "loss": 3.746735, "lr_stage": "cosine_decay", "mfu": 58.170377, "stage_name": "permutation_cycle", "status": "training", "step": 294, "tokens_per_sec_per_gpu": 101264.753255, "vram_gb": 20.709735} {"adam_lr": 9.931428007686158e-05, "dt_ms": 161.990404, "global_tokens_per_sec": 101141.793479, "grad_norm": 42.25, "loss": 2.96924, "lr_stage": "cosine_decay", "mfu": 58.099744, "stage_name": "permutation_cycle", "status": "training", "step": 295, "tokens_per_sec_per_gpu": 101141.793479, "vram_gb": 20.709735} {"adam_lr": 9.929980185352526e-05, "dt_ms": 163.640499, "global_tokens_per_sec": 100121.914126, "grad_norm": 54.25, "loss": 3.722705, "lr_stage": "cosine_decay", "mfu": 57.513886, "stage_name": "permutation_cycle", "status": "training", "step": 296, "tokens_per_sec_per_gpu": 100121.914126, "vram_gb": 20.709735} {"adam_lr": 9.928517345444269e-05, "dt_ms": 164.642811, "global_tokens_per_sec": 99512.392422, "grad_norm": 42.0, "loss": 4.096201, "lr_stage": "cosine_decay", "mfu": 57.163754, "stage_name": "permutation_cycle", "status": "training", "step": 297, "tokens_per_sec_per_gpu": 99512.392422, "vram_gb": 20.709735} {"adam_lr": 9.927039492417452e-05, "dt_ms": 162.752867, "global_tokens_per_sec": 100667.965657, "grad_norm": 42.5, "loss": 3.69359, "lr_stage": "cosine_decay", "mfu": 57.827559, "stage_name": "permutation_cycle", "status": "training", "step": 298, "tokens_per_sec_per_gpu": 100667.965657, "vram_gb": 20.709735} {"adam_lr": 9.92554663077387e-05, "dt_ms": 163.540602, "global_tokens_per_sec": 100183.072746, "grad_norm": 52.0, "loss": 3.605868, "lr_stage": "cosine_decay", "mfu": 57.549018, "stage_name": "permutation_cycle", "status": "training", "step": 299, "tokens_per_sec_per_gpu": 100183.072746, "vram_gb": 20.709735} {"adam_lr": 9.924038765061042e-05, "dt_ms": 162.14323, "global_tokens_per_sec": 101046.4634, "grad_norm": 44.75, "loss": 2.737271, "lr_stage": "cosine_decay", "mfu": 58.044983, "stage_name": "permutation_cycle", "status": "training", "step": 300, "tokens_per_sec_per_gpu": 101046.4634, "vram_gb": 20.709735} {"adam_lr": 9.922515899872184e-05, "dt_ms": 161.913157, "global_tokens_per_sec": 101190.047512, "grad_norm": 49.0, "loss": 4.940782, "lr_stage": "cosine_decay", "mfu": 58.127463, "stage_name": "permutation_cycle", "status": "training", "step": 301, "tokens_per_sec_per_gpu": 101190.047512, "vram_gb": 20.709735} {"adam_lr": 9.92097803984621e-05, "dt_ms": 161.818743, "global_tokens_per_sec": 101249.087228, "grad_norm": 46.5, "loss": 2.924186, "lr_stage": "cosine_decay", "mfu": 58.161378, "stage_name": "permutation_cycle", "status": "training", "step": 302, "tokens_per_sec_per_gpu": 101249.087228, "vram_gb": 20.709735} {"adam_lr": 9.919425189667709e-05, "dt_ms": 164.041519, "global_tokens_per_sec": 99877.153561, "grad_norm": 37.0, "loss": 2.217531, "lr_stage": "cosine_decay", "mfu": 57.373286, "stage_name": "permutation_cycle", "status": "training", "step": 303, "tokens_per_sec_per_gpu": 99877.153561, "vram_gb": 20.709735} {"adam_lr": 9.917857354066931e-05, "dt_ms": 164.235353, "global_tokens_per_sec": 99759.276269, "grad_norm": 62.25, "loss": 4.281571, "lr_stage": "cosine_decay", "mfu": 57.305573, "stage_name": "permutation_cycle", "status": "training", "step": 304, "tokens_per_sec_per_gpu": 99759.276269, "vram_gb": 20.709735} {"adam_lr": 9.916274537819775e-05, "dt_ms": 163.335323, "global_tokens_per_sec": 100308.98195, "grad_norm": 47.5, "loss": 3.854045, "lr_stage": "cosine_decay", "mfu": 57.621345, "stage_name": "permutation_cycle", "status": "training", "step": 305, "tokens_per_sec_per_gpu": 100308.98195, "vram_gb": 20.709735} {"adam_lr": 9.914676745747772e-05, "dt_ms": 163.521051, "global_tokens_per_sec": 100195.050478, "grad_norm": 35.25, "loss": 4.114156, "lr_stage": "cosine_decay", "mfu": 57.555899, "stage_name": "permutation_cycle", "status": "training", "step": 306, "tokens_per_sec_per_gpu": 100195.050478, "vram_gb": 20.709735} {"adam_lr": 9.913063982718076e-05, "dt_ms": 163.340807, "global_tokens_per_sec": 100305.614407, "grad_norm": 35.25, "loss": 4.019533, "lr_stage": "cosine_decay", "mfu": 57.619411, "stage_name": "permutation_cycle", "status": "training", "step": 307, "tokens_per_sec_per_gpu": 100305.614407, "vram_gb": 20.709735} {"adam_lr": 9.911436253643445e-05, "dt_ms": 162.133694, "global_tokens_per_sec": 101052.406977, "grad_norm": 60.0, "loss": 4.02714, "lr_stage": "cosine_decay", "mfu": 58.048397, "stage_name": "permutation_cycle", "status": "training", "step": 308, "tokens_per_sec_per_gpu": 101052.406977, "vram_gb": 20.709735} {"adam_lr": 9.909793563482219e-05, "dt_ms": 164.897203, "global_tokens_per_sec": 99358.87121, "grad_norm": 48.25, "loss": 4.798974, "lr_stage": "cosine_decay", "mfu": 57.075565, "stage_name": "permutation_cycle", "status": "training", "step": 309, "tokens_per_sec_per_gpu": 99358.87121, "vram_gb": 20.709735} {"adam_lr": 9.908135917238321e-05, "dt_ms": 164.158583, "global_tokens_per_sec": 99805.929923, "grad_norm": 37.0, "loss": 2.976363, "lr_stage": "cosine_decay", "mfu": 57.332373, "stage_name": "permutation_cycle", "status": "training", "step": 310, "tokens_per_sec_per_gpu": 99805.929923, "vram_gb": 20.709735} {"adam_lr": 9.906463319961225e-05, "dt_ms": 163.83481, "global_tokens_per_sec": 100003.167668, "grad_norm": 47.5, "loss": 3.97235, "lr_stage": "cosine_decay", "mfu": 57.445674, "stage_name": "permutation_cycle", "status": "training", "step": 311, "tokens_per_sec_per_gpu": 100003.167668, "vram_gb": 20.709735} {"adam_lr": 9.904775776745958e-05, "dt_ms": 161.300898, "global_tokens_per_sec": 101574.14028, "grad_norm": 38.25, "loss": 2.958781, "lr_stage": "cosine_decay", "mfu": 58.348101, "stage_name": "permutation_cycle", "status": "training", "step": 312, "tokens_per_sec_per_gpu": 101574.14028, "vram_gb": 20.709735} {"adam_lr": 9.903073292733065e-05, "dt_ms": 162.248135, "global_tokens_per_sec": 100981.130163, "grad_norm": 31.75, "loss": 3.572306, "lr_stage": "cosine_decay", "mfu": 58.007453, "stage_name": "permutation_cycle", "status": "training", "step": 313, "tokens_per_sec_per_gpu": 100981.130163, "vram_gb": 20.709735} {"adam_lr": 9.901355873108609e-05, "dt_ms": 163.220882, "global_tokens_per_sec": 100379.31273, "grad_norm": 54.5, "loss": 4.193754, "lr_stage": "cosine_decay", "mfu": 57.661746, "stage_name": "permutation_cycle", "status": "training", "step": 314, "tokens_per_sec_per_gpu": 100379.31273, "vram_gb": 20.709735} {"adam_lr": 9.899623523104149e-05, "dt_ms": 163.287878, "global_tokens_per_sec": 100338.127955, "grad_norm": 34.75, "loss": 2.278263, "lr_stage": "cosine_decay", "mfu": 57.638088, "stage_name": "permutation_cycle", "status": "training", "step": 315, "tokens_per_sec_per_gpu": 100338.127955, "vram_gb": 20.709735} {"adam_lr": 9.89787624799672e-05, "dt_ms": 165.081978, "global_tokens_per_sec": 99247.659944, "grad_norm": 72.5, "loss": 4.305115, "lr_stage": "cosine_decay", "mfu": 57.011681, "stage_name": "permutation_cycle", "status": "training", "step": 316, "tokens_per_sec_per_gpu": 99247.659944, "vram_gb": 20.709735} {"adam_lr": 9.896114053108829e-05, "dt_ms": 165.002346, "global_tokens_per_sec": 99295.557871, "grad_norm": 34.25, "loss": 4.427789, "lr_stage": "cosine_decay", "mfu": 57.039196, "stage_name": "permutation_cycle", "status": "training", "step": 317, "tokens_per_sec_per_gpu": 99295.557871, "vram_gb": 20.709735} {"adam_lr": 9.894336943808426e-05, "dt_ms": 163.615942, "global_tokens_per_sec": 100136.941423, "grad_norm": 55.0, "loss": 3.713417, "lr_stage": "cosine_decay", "mfu": 57.522519, "stage_name": "permutation_cycle", "status": "training", "step": 318, "tokens_per_sec_per_gpu": 100136.941423, "vram_gb": 20.709735} {"adam_lr": 9.892544925508893e-05, "dt_ms": 161.126137, "global_tokens_per_sec": 101684.309743, "grad_norm": 53.75, "loss": 4.87566, "lr_stage": "cosine_decay", "mfu": 58.411387, "stage_name": "permutation_cycle", "status": "training", "step": 319, "tokens_per_sec_per_gpu": 101684.309743, "vram_gb": 20.709735} {"adam_lr": 9.890738003669029e-05, "dt_ms": 161.815166, "global_tokens_per_sec": 101251.324935, "grad_norm": 42.0, "loss": 3.431933, "lr_stage": "cosine_decay", "mfu": 58.162664, "stage_name": "permutation_cycle", "status": "training", "step": 320, "tokens_per_sec_per_gpu": 101251.324935, "vram_gb": 20.709735} {"adam_lr": 9.888916183793031e-05, "dt_ms": 162.605762, "global_tokens_per_sec": 100759.036764, "grad_norm": 43.25, "loss": 3.8272, "lr_stage": "cosine_decay", "mfu": 57.879874, "stage_name": "permutation_cycle", "status": "training", "step": 321, "tokens_per_sec_per_gpu": 100759.036764, "vram_gb": 20.709735} {"adam_lr": 9.88707947143048e-05, "dt_ms": 163.128376, "global_tokens_per_sec": 100436.235565, "grad_norm": 69.0, "loss": 3.705659, "lr_stage": "cosine_decay", "mfu": 57.694445, "stage_name": "permutation_cycle", "status": "training", "step": 322, "tokens_per_sec_per_gpu": 100436.235565, "vram_gb": 20.709735} {"adam_lr": 9.88522787217632e-05, "dt_ms": 163.996696, "global_tokens_per_sec": 99904.451446, "grad_norm": 45.5, "loss": 4.28826, "lr_stage": "cosine_decay", "mfu": 57.388967, "stage_name": "permutation_cycle", "status": "training", "step": 323, "tokens_per_sec_per_gpu": 99904.451446, "vram_gb": 20.709735} {"adam_lr": 9.88336139167084e-05, "dt_ms": 165.551901, "global_tokens_per_sec": 98965.943094, "grad_norm": 41.25, "loss": 3.797483, "lr_stage": "cosine_decay", "mfu": 56.849852, "stage_name": "permutation_cycle", "status": "training", "step": 324, "tokens_per_sec_per_gpu": 98965.943094, "vram_gb": 20.709735} {"adam_lr": 9.881480035599667e-05, "dt_ms": 163.003206, "global_tokens_per_sec": 100513.360299, "grad_norm": 44.5, "loss": 3.364148, "lr_stage": "cosine_decay", "mfu": 57.738748, "stage_name": "permutation_cycle", "status": "training", "step": 325, "tokens_per_sec_per_gpu": 100513.360299, "vram_gb": 20.709735} {"adam_lr": 9.879583809693738e-05, "dt_ms": 163.465738, "global_tokens_per_sec": 100228.954218, "grad_norm": 57.75, "loss": 3.72997, "lr_stage": "cosine_decay", "mfu": 57.575374, "stage_name": "permutation_cycle", "status": "training", "step": 326, "tokens_per_sec_per_gpu": 100228.954218, "vram_gb": 20.709735} {"adam_lr": 9.877672719729282e-05, "dt_ms": 162.610292, "global_tokens_per_sec": 100756.229847, "grad_norm": 47.25, "loss": 3.685045, "lr_stage": "cosine_decay", "mfu": 57.878262, "stage_name": "permutation_cycle", "status": "training", "step": 327, "tokens_per_sec_per_gpu": 100756.229847, "vram_gb": 20.709735} {"adam_lr": 9.875746771527816e-05, "dt_ms": 163.467407, "global_tokens_per_sec": 100227.930925, "grad_norm": 40.0, "loss": 3.492059, "lr_stage": "cosine_decay", "mfu": 57.574787, "stage_name": "permutation_cycle", "status": "training", "step": 328, "tokens_per_sec_per_gpu": 100227.930925, "vram_gb": 20.709735} {"adam_lr": 9.87380597095611e-05, "dt_ms": 161.742926, "global_tokens_per_sec": 101296.547807, "grad_norm": 40.75, "loss": 3.736194, "lr_stage": "cosine_decay", "mfu": 58.188641, "stage_name": "permutation_cycle", "status": "training", "step": 329, "tokens_per_sec_per_gpu": 101296.547807, "vram_gb": 20.709735} {"adam_lr": 9.871850323926177e-05, "dt_ms": 164.885521, "global_tokens_per_sec": 99365.91101, "grad_norm": 43.75, "loss": 3.734581, "lr_stage": "cosine_decay", "mfu": 57.079609, "stage_name": "permutation_cycle", "status": "training", "step": 330, "tokens_per_sec_per_gpu": 99365.91101, "vram_gb": 20.709735} {"adam_lr": 9.86987983639526e-05, "dt_ms": 162.966967, "global_tokens_per_sec": 100535.711862, "grad_norm": 31.125, "loss": 3.384903, "lr_stage": "cosine_decay", "mfu": 57.751588, "stage_name": "permutation_cycle", "status": "training", "step": 331, "tokens_per_sec_per_gpu": 100535.711862, "vram_gb": 20.709735} {"adam_lr": 9.867894514365802e-05, "dt_ms": 162.425041, "global_tokens_per_sec": 100871.145724, "grad_norm": 44.0, "loss": 3.57641, "lr_stage": "cosine_decay", "mfu": 57.944274, "stage_name": "permutation_cycle", "status": "training", "step": 332, "tokens_per_sec_per_gpu": 100871.145724, "vram_gb": 20.709735} {"adam_lr": 9.865894363885442e-05, "dt_ms": 162.662029, "global_tokens_per_sec": 100724.182982, "grad_norm": 63.0, "loss": 4.048018, "lr_stage": "cosine_decay", "mfu": 57.859853, "stage_name": "permutation_cycle", "status": "training", "step": 333, "tokens_per_sec_per_gpu": 100724.182982, "vram_gb": 20.709735} {"adam_lr": 9.863879391046984e-05, "dt_ms": 163.44595, "global_tokens_per_sec": 100241.089147, "grad_norm": 37.5, "loss": 4.45252, "lr_stage": "cosine_decay", "mfu": 57.582345, "stage_name": "permutation_cycle", "status": "training", "step": 334, "tokens_per_sec_per_gpu": 100241.089147, "vram_gb": 20.709735} {"adam_lr": 9.861849601988383e-05, "dt_ms": 163.347721, "global_tokens_per_sec": 100301.368698, "grad_norm": 40.25, "loss": 3.26218, "lr_stage": "cosine_decay", "mfu": 57.616972, "stage_name": "permutation_cycle", "status": "training", "step": 335, "tokens_per_sec_per_gpu": 100301.368698, "vram_gb": 20.709735} {"adam_lr": 9.859805002892732e-05, "dt_ms": 163.428068, "global_tokens_per_sec": 100252.056971, "grad_norm": 53.25, "loss": 4.127469, "lr_stage": "cosine_decay", "mfu": 57.588645, "stage_name": "permutation_cycle", "status": "training", "step": 336, "tokens_per_sec_per_gpu": 100252.056971, "vram_gb": 20.709735} {"adam_lr": 9.85774559998823e-05, "dt_ms": 163.496971, "global_tokens_per_sec": 100209.807477, "grad_norm": 45.25, "loss": 4.539279, "lr_stage": "cosine_decay", "mfu": 57.564376, "stage_name": "permutation_cycle", "status": "training", "step": 337, "tokens_per_sec_per_gpu": 100209.807477, "vram_gb": 20.709735} {"adam_lr": 9.855671399548181e-05, "dt_ms": 163.070202, "global_tokens_per_sec": 100472.065477, "grad_norm": 35.25, "loss": 3.831316, "lr_stage": "cosine_decay", "mfu": 57.715027, "stage_name": "permutation_cycle", "status": "training", "step": 338, "tokens_per_sec_per_gpu": 100472.065477, "vram_gb": 20.709735} {"adam_lr": 9.853582407890954e-05, "dt_ms": 162.705421, "global_tokens_per_sec": 100697.320681, "grad_norm": 40.5, "loss": 3.072792, "lr_stage": "cosine_decay", "mfu": 57.844422, "stage_name": "permutation_cycle", "status": "training", "step": 339, "tokens_per_sec_per_gpu": 100697.320681, "vram_gb": 20.709735} {"adam_lr": 9.851478631379982e-05, "dt_ms": 161.278486, "global_tokens_per_sec": 101588.255078, "grad_norm": 65.0, "loss": 2.570343, "lr_stage": "cosine_decay", "mfu": 58.356209, "stage_name": "permutation_cycle", "status": "training", "step": 340, "tokens_per_sec_per_gpu": 101588.255078, "vram_gb": 20.709735} {"adam_lr": 9.849360076423734e-05, "dt_ms": 164.413691, "global_tokens_per_sec": 99651.068859, "grad_norm": 53.0, "loss": 3.608577, "lr_stage": "cosine_decay", "mfu": 57.243415, "stage_name": "permutation_cycle", "status": "training", "step": 341, "tokens_per_sec_per_gpu": 99651.068859, "vram_gb": 20.709735} {"adam_lr": 9.847226749475695e-05, "dt_ms": 163.19108, "global_tokens_per_sec": 100397.644226, "grad_norm": 54.5, "loss": 3.590718, "lr_stage": "cosine_decay", "mfu": 57.672276, "stage_name": "permutation_cycle", "status": "training", "step": 342, "tokens_per_sec_per_gpu": 100397.644226, "vram_gb": 20.709735} {"adam_lr": 9.845078657034348e-05, "dt_ms": 163.293839, "global_tokens_per_sec": 100334.465467, "grad_norm": 44.75, "loss": 3.660611, "lr_stage": "cosine_decay", "mfu": 57.635984, "stage_name": "permutation_cycle", "status": "training", "step": 343, "tokens_per_sec_per_gpu": 100334.465467, "vram_gb": 20.709735} {"adam_lr": 9.842915805643155e-05, "dt_ms": 163.217545, "global_tokens_per_sec": 100381.365524, "grad_norm": 47.0, "loss": 4.296735, "lr_stage": "cosine_decay", "mfu": 57.662925, "stage_name": "permutation_cycle", "status": "training", "step": 344, "tokens_per_sec_per_gpu": 100381.365524, "vram_gb": 20.709735} {"adam_lr": 9.84073820189054e-05, "dt_ms": 164.095879, "global_tokens_per_sec": 99844.067625, "grad_norm": 32.5, "loss": 3.716637, "lr_stage": "cosine_decay", "mfu": 57.354281, "stage_name": "permutation_cycle", "status": "training", "step": 345, "tokens_per_sec_per_gpu": 99844.067625, "vram_gb": 20.709735} {"adam_lr": 9.838545852409857e-05, "dt_ms": 162.813425, "global_tokens_per_sec": 100630.52229, "grad_norm": 35.0, "loss": 3.274621, "lr_stage": "cosine_decay", "mfu": 57.806051, "stage_name": "permutation_cycle", "status": "training", "step": 346, "tokens_per_sec_per_gpu": 100630.52229, "vram_gb": 20.709735} {"adam_lr": 9.836338763879385e-05, "dt_ms": 162.487268, "global_tokens_per_sec": 100832.515412, "grad_norm": 60.75, "loss": 4.404417, "lr_stage": "cosine_decay", "mfu": 57.922083, "stage_name": "permutation_cycle", "status": "training", "step": 347, "tokens_per_sec_per_gpu": 100832.515412, "vram_gb": 20.709735} {"adam_lr": 9.834116943022298e-05, "dt_ms": 165.40432, "global_tokens_per_sec": 99054.244916, "grad_norm": 33.0, "loss": 3.162657, "lr_stage": "cosine_decay", "mfu": 56.900576, "stage_name": "permutation_cycle", "status": "training", "step": 348, "tokens_per_sec_per_gpu": 99054.244916, "vram_gb": 20.709735} {"adam_lr": 9.831880396606648e-05, "dt_ms": 161.215305, "global_tokens_per_sec": 101628.067922, "grad_norm": 39.5, "loss": 4.087339, "lr_stage": "cosine_decay", "mfu": 58.379079, "stage_name": "permutation_cycle", "status": "training", "step": 349, "tokens_per_sec_per_gpu": 101628.067922, "vram_gb": 20.709735} {"adam_lr": 9.829629131445342e-05, "dt_ms": 162.175894, "global_tokens_per_sec": 101026.111944, "grad_norm": 35.0, "loss": 3.983428, "lr_stage": "cosine_decay", "mfu": 58.033292, "stage_name": "permutation_cycle", "status": "training", "step": 350, "tokens_per_sec_per_gpu": 101026.111944, "vram_gb": 20.709735} {"adam_lr": 9.827363154396125e-05, "dt_ms": 164.03842, "global_tokens_per_sec": 99879.0407, "grad_norm": 41.5, "loss": 3.868796, "lr_stage": "cosine_decay", "mfu": 57.374371, "stage_name": "permutation_cycle", "status": "training", "step": 351, "tokens_per_sec_per_gpu": 99879.0407, "vram_gb": 20.709735} {"adam_lr": 9.825082472361557e-05, "dt_ms": 164.226294, "global_tokens_per_sec": 99764.77971, "grad_norm": 50.75, "loss": 3.72958, "lr_stage": "cosine_decay", "mfu": 57.308735, "stage_name": "permutation_cycle", "status": "training", "step": 352, "tokens_per_sec_per_gpu": 99764.77971, "vram_gb": 20.709735} {"adam_lr": 9.822787092288991e-05, "dt_ms": 162.427187, "global_tokens_per_sec": 100869.813152, "grad_norm": 35.25, "loss": 3.097893, "lr_stage": "cosine_decay", "mfu": 57.943508, "stage_name": "permutation_cycle", "status": "training", "step": 353, "tokens_per_sec_per_gpu": 100869.813152, "vram_gb": 20.709735} {"adam_lr": 9.820477021170551e-05, "dt_ms": 165.225744, "global_tokens_per_sec": 99161.302463, "grad_norm": 46.75, "loss": 4.785278, "lr_stage": "cosine_decay", "mfu": 56.962074, "stage_name": "permutation_cycle", "status": "training", "step": 354, "tokens_per_sec_per_gpu": 99161.302463, "vram_gb": 20.709735} {"adam_lr": 9.818152266043114e-05, "dt_ms": 163.258314, "global_tokens_per_sec": 100356.297853, "grad_norm": 42.75, "loss": 3.939732, "lr_stage": "cosine_decay", "mfu": 57.648525, "stage_name": "permutation_cycle", "status": "training", "step": 355, "tokens_per_sec_per_gpu": 100356.297853, "vram_gb": 20.709735} {"adam_lr": 9.815812833988291e-05, "dt_ms": 161.644936, "global_tokens_per_sec": 101357.954324, "grad_norm": 52.75, "loss": 4.005147, "lr_stage": "cosine_decay", "mfu": 58.223916, "stage_name": "permutation_cycle", "status": "training", "step": 356, "tokens_per_sec_per_gpu": 101357.954324, "vram_gb": 20.709735} {"adam_lr": 9.813458732132395e-05, "dt_ms": 160.674334, "global_tokens_per_sec": 101970.237783, "grad_norm": 46.5, "loss": 3.336308, "lr_stage": "cosine_decay", "mfu": 58.575635, "stage_name": "permutation_cycle", "status": "training", "step": 357, "tokens_per_sec_per_gpu": 101970.237783, "vram_gb": 20.709735} {"adam_lr": 9.811089967646428e-05, "dt_ms": 162.293911, "global_tokens_per_sec": 100952.647583, "grad_norm": 51.0, "loss": 3.225451, "lr_stage": "cosine_decay", "mfu": 57.991092, "stage_name": "permutation_cycle", "status": "training", "step": 358, "tokens_per_sec_per_gpu": 100952.647583, "vram_gb": 20.709735} {"adam_lr": 9.808706547746057e-05, "dt_ms": 164.064169, "global_tokens_per_sec": 99863.36509, "grad_norm": 34.25, "loss": 2.472488, "lr_stage": "cosine_decay", "mfu": 57.365366, "stage_name": "permutation_cycle", "status": "training", "step": 359, "tokens_per_sec_per_gpu": 99863.36509, "vram_gb": 20.709735} {"adam_lr": 9.806308479691595e-05, "dt_ms": 164.489746, "global_tokens_per_sec": 99604.99295, "grad_norm": 47.25, "loss": 3.722103, "lr_stage": "cosine_decay", "mfu": 57.216947, "stage_name": "permutation_cycle", "status": "training", "step": 360, "tokens_per_sec_per_gpu": 99604.99295, "vram_gb": 20.709735} {"adam_lr": 9.803895770787972e-05, "dt_ms": 162.612915, "global_tokens_per_sec": 100754.604861, "grad_norm": 37.0, "loss": 3.162034, "lr_stage": "cosine_decay", "mfu": 57.877328, "stage_name": "permutation_cycle", "status": "training", "step": 361, "tokens_per_sec_per_gpu": 100754.604861, "vram_gb": 20.709735} {"adam_lr": 9.801468428384716e-05, "dt_ms": 162.766695, "global_tokens_per_sec": 100659.413142, "grad_norm": 30.125, "loss": 3.251388, "lr_stage": "cosine_decay", "mfu": 57.822647, "stage_name": "permutation_cycle", "status": "training", "step": 362, "tokens_per_sec_per_gpu": 100659.413142, "vram_gb": 20.709735} {"adam_lr": 9.799026459875935e-05, "dt_ms": 162.128925, "global_tokens_per_sec": 101055.379028, "grad_norm": 40.25, "loss": 2.863745, "lr_stage": "cosine_decay", "mfu": 58.050105, "stage_name": "permutation_cycle", "status": "training", "step": 363, "tokens_per_sec_per_gpu": 101055.379028, "vram_gb": 20.709735} {"adam_lr": 9.796569872700288e-05, "dt_ms": 162.971258, "global_tokens_per_sec": 100533.064447, "grad_norm": 41.25, "loss": 3.556285, "lr_stage": "cosine_decay", "mfu": 57.750067, "stage_name": "permutation_cycle", "status": "training", "step": 364, "tokens_per_sec_per_gpu": 100533.064447, "vram_gb": 20.709735} {"adam_lr": 9.794098674340965e-05, "dt_ms": 163.765907, "global_tokens_per_sec": 100045.243063, "grad_norm": 40.75, "loss": 2.632606, "lr_stage": "cosine_decay", "mfu": 57.469844, "stage_name": "permutation_cycle", "status": "training", "step": 365, "tokens_per_sec_per_gpu": 100045.243063, "vram_gb": 20.709735} {"adam_lr": 9.791612872325667e-05, "dt_ms": 165.130377, "global_tokens_per_sec": 99218.570901, "grad_norm": 35.75, "loss": 4.118549, "lr_stage": "cosine_decay", "mfu": 56.994971, "stage_name": "permutation_cycle", "status": "training", "step": 366, "tokens_per_sec_per_gpu": 99218.570901, "vram_gb": 20.709735} {"adam_lr": 9.789112474226575e-05, "dt_ms": 161.808968, "global_tokens_per_sec": 101255.203862, "grad_norm": 39.5, "loss": 3.766646, "lr_stage": "cosine_decay", "mfu": 58.164892, "stage_name": "permutation_cycle", "status": "training", "step": 367, "tokens_per_sec_per_gpu": 101255.203862, "vram_gb": 20.709735} {"adam_lr": 9.786597487660337e-05, "dt_ms": 162.525654, "global_tokens_per_sec": 100808.700737, "grad_norm": 31.375, "loss": 2.656901, "lr_stage": "cosine_decay", "mfu": 57.908403, "stage_name": "permutation_cycle", "status": "training", "step": 368, "tokens_per_sec_per_gpu": 100808.700737, "vram_gb": 20.709735} {"adam_lr": 9.784067920288038e-05, "dt_ms": 162.917137, "global_tokens_per_sec": 100566.461497, "grad_norm": 53.75, "loss": 4.126459, "lr_stage": "cosine_decay", "mfu": 57.769252, "stage_name": "permutation_cycle", "status": "training", "step": 369, "tokens_per_sec_per_gpu": 100566.461497, "vram_gb": 20.709735} {"adam_lr": 9.781523779815179e-05, "dt_ms": 161.871433, "global_tokens_per_sec": 101216.129803, "grad_norm": 38.75, "loss": 2.594083, "lr_stage": "cosine_decay", "mfu": 58.142446, "stage_name": "permutation_cycle", "status": "training", "step": 370, "tokens_per_sec_per_gpu": 101216.129803, "vram_gb": 20.709735} {"adam_lr": 9.778965073991651e-05, "dt_ms": 163.300991, "global_tokens_per_sec": 100330.070833, "grad_norm": 44.0, "loss": 2.996114, "lr_stage": "cosine_decay", "mfu": 57.63346, "stage_name": "permutation_cycle", "status": "training", "step": 371, "tokens_per_sec_per_gpu": 100330.070833, "vram_gb": 20.709735} {"adam_lr": 9.776391810611718e-05, "dt_ms": 163.81073, "global_tokens_per_sec": 100017.868194, "grad_norm": 47.75, "loss": 2.744751, "lr_stage": "cosine_decay", "mfu": 57.454118, "stage_name": "permutation_cycle", "status": "training", "step": 372, "tokens_per_sec_per_gpu": 100017.868194, "vram_gb": 20.709735} {"adam_lr": 9.773803997513989e-05, "dt_ms": 161.202431, "global_tokens_per_sec": 101636.184556, "grad_norm": 49.25, "loss": 3.664136, "lr_stage": "cosine_decay", "mfu": 58.383742, "stage_name": "permutation_cycle", "status": "training", "step": 373, "tokens_per_sec_per_gpu": 101636.184556, "vram_gb": 20.709735} {"adam_lr": 9.771201642581385e-05, "dt_ms": 162.585974, "global_tokens_per_sec": 100771.300397, "grad_norm": 36.25, "loss": 3.233593, "lr_stage": "cosine_decay", "mfu": 57.886919, "stage_name": "permutation_cycle", "status": "training", "step": 374, "tokens_per_sec_per_gpu": 100771.300397, "vram_gb": 20.709735} {"adam_lr": 9.768584753741134e-05, "dt_ms": 161.982059, "global_tokens_per_sec": 101147.003889, "grad_norm": 40.5, "loss": 3.845143, "lr_stage": "cosine_decay", "mfu": 58.102737, "stage_name": "permutation_cycle", "status": "training", "step": 375, "tokens_per_sec_per_gpu": 101147.003889, "vram_gb": 20.709735} {"adam_lr": 9.765953338964735e-05, "dt_ms": 161.894798, "global_tokens_per_sec": 101201.522064, "grad_norm": 41.5, "loss": 3.939978, "lr_stage": "cosine_decay", "mfu": 58.134055, "stage_name": "permutation_cycle", "status": "training", "step": 376, "tokens_per_sec_per_gpu": 101201.522064, "vram_gb": 20.709735} {"adam_lr": 9.763307406267932e-05, "dt_ms": 163.162231, "global_tokens_per_sec": 100415.395492, "grad_norm": 36.0, "loss": 2.692783, "lr_stage": "cosine_decay", "mfu": 57.682473, "stage_name": "permutation_cycle", "status": "training", "step": 377, "tokens_per_sec_per_gpu": 100415.395492, "vram_gb": 20.709735} {"adam_lr": 9.760646963710694e-05, "dt_ms": 164.086819, "global_tokens_per_sec": 99849.580425, "grad_norm": 35.75, "loss": 2.330479, "lr_stage": "cosine_decay", "mfu": 57.357447, "stage_name": "permutation_cycle", "status": "training", "step": 378, "tokens_per_sec_per_gpu": 99849.580425, "vram_gb": 20.709735} {"adam_lr": 9.757972019397192e-05, "dt_ms": 162.553549, "global_tokens_per_sec": 100791.401478, "grad_norm": 50.0, "loss": 3.85493, "lr_stage": "cosine_decay", "mfu": 57.898466, "stage_name": "permutation_cycle", "status": "training", "step": 379, "tokens_per_sec_per_gpu": 100791.401478, "vram_gb": 20.709735} {"adam_lr": 9.755282581475769e-05, "dt_ms": 162.999153, "global_tokens_per_sec": 100515.859651, "grad_norm": 43.5, "loss": 3.87128, "lr_stage": "cosine_decay", "mfu": 57.740184, "stage_name": "permutation_cycle", "status": "training", "step": 380, "tokens_per_sec_per_gpu": 100515.859651, "vram_gb": 20.709735} {"adam_lr": 9.752578658138919e-05, "dt_ms": 162.948132, "global_tokens_per_sec": 100547.332719, "grad_norm": 34.25, "loss": 2.163344, "lr_stage": "cosine_decay", "mfu": 57.758263, "stage_name": "permutation_cycle", "status": "training", "step": 381, "tokens_per_sec_per_gpu": 100547.332719, "vram_gb": 20.709735} {"adam_lr": 9.749860257623263e-05, "dt_ms": 162.528276, "global_tokens_per_sec": 100807.074058, "grad_norm": 43.75, "loss": 3.741748, "lr_stage": "cosine_decay", "mfu": 57.907469, "stage_name": "permutation_cycle", "status": "training", "step": 382, "tokens_per_sec_per_gpu": 100807.074058, "vram_gb": 20.709735} {"adam_lr": 9.74712738820952e-05, "dt_ms": 164.251328, "global_tokens_per_sec": 99749.574313, "grad_norm": 35.0, "loss": 3.887065, "lr_stage": "cosine_decay", "mfu": 57.3, "stage_name": "permutation_cycle", "status": "training", "step": 383, "tokens_per_sec_per_gpu": 99749.574313, "vram_gb": 20.709735} {"adam_lr": 9.744380058222483e-05, "dt_ms": 163.106441, "global_tokens_per_sec": 100449.742202, "grad_norm": 46.25, "loss": 3.675487, "lr_stage": "cosine_decay", "mfu": 57.702203, "stage_name": "permutation_cycle", "status": "training", "step": 384, "tokens_per_sec_per_gpu": 100449.742202, "vram_gb": 20.709735} {"adam_lr": 9.741618276030997e-05, "dt_ms": 162.169218, "global_tokens_per_sec": 101030.270699, "grad_norm": 36.0, "loss": 3.316943, "lr_stage": "cosine_decay", "mfu": 58.035681, "stage_name": "permutation_cycle", "status": "training", "step": 385, "tokens_per_sec_per_gpu": 101030.270699, "vram_gb": 20.709735} {"adam_lr": 9.73884205004793e-05, "dt_ms": 162.365675, "global_tokens_per_sec": 100908.02753, "grad_norm": 51.75, "loss": 3.499694, "lr_stage": "cosine_decay", "mfu": 57.96546, "stage_name": "permutation_cycle", "status": "training", "step": 386, "tokens_per_sec_per_gpu": 100908.02753, "vram_gb": 20.709735} {"adam_lr": 9.736051388730145e-05, "dt_ms": 162.939072, "global_tokens_per_sec": 100552.923455, "grad_norm": 30.5, "loss": 3.771019, "lr_stage": "cosine_decay", "mfu": 57.761475, "stage_name": "permutation_cycle", "status": "training", "step": 387, "tokens_per_sec_per_gpu": 100552.923455, "vram_gb": 20.709735} {"adam_lr": 9.733246300578483e-05, "dt_ms": 163.681746, "global_tokens_per_sec": 100096.684252, "grad_norm": 39.25, "loss": 3.295321, "lr_stage": "cosine_decay", "mfu": 57.499393, "stage_name": "permutation_cycle", "status": "training", "step": 388, "tokens_per_sec_per_gpu": 100096.684252, "vram_gb": 20.709735} {"adam_lr": 9.730426794137727e-05, "dt_ms": 162.754059, "global_tokens_per_sec": 100667.228314, "grad_norm": 47.75, "loss": 3.737393, "lr_stage": "cosine_decay", "mfu": 57.827136, "stage_name": "permutation_cycle", "status": "training", "step": 389, "tokens_per_sec_per_gpu": 100667.228314, "vram_gb": 20.709735} {"adam_lr": 9.727592877996585e-05, "dt_ms": 164.111376, "global_tokens_per_sec": 99834.639246, "grad_norm": 46.75, "loss": 4.08246, "lr_stage": "cosine_decay", "mfu": 57.348865, "stage_name": "permutation_cycle", "status": "training", "step": 390, "tokens_per_sec_per_gpu": 99834.639246, "vram_gb": 20.709735} {"adam_lr": 9.724744560787655e-05, "dt_ms": 163.312435, "global_tokens_per_sec": 100323.04022, "grad_norm": 34.25, "loss": 3.503431, "lr_stage": "cosine_decay", "mfu": 57.629421, "stage_name": "permutation_cycle", "status": "training", "step": 391, "tokens_per_sec_per_gpu": 100323.04022, "vram_gb": 20.709735} {"adam_lr": 9.721881851187406e-05, "dt_ms": 161.389828, "global_tokens_per_sec": 101518.170201, "grad_norm": 51.5, "loss": 3.122831, "lr_stage": "cosine_decay", "mfu": 58.31595, "stage_name": "permutation_cycle", "status": "training", "step": 392, "tokens_per_sec_per_gpu": 101518.170201, "vram_gb": 20.709735} {"adam_lr": 9.719004757916149e-05, "dt_ms": 164.527178, "global_tokens_per_sec": 99582.331734, "grad_norm": 54.75, "loss": 4.462381, "lr_stage": "cosine_decay", "mfu": 57.203929, "stage_name": "permutation_cycle", "status": "training", "step": 393, "tokens_per_sec_per_gpu": 99582.331734, "vram_gb": 20.709735} {"adam_lr": 9.716113289738004e-05, "dt_ms": 163.56492, "global_tokens_per_sec": 100168.17761, "grad_norm": 33.25, "loss": 3.133618, "lr_stage": "cosine_decay", "mfu": 57.540462, "stage_name": "permutation_cycle", "status": "training", "step": 394, "tokens_per_sec_per_gpu": 100168.17761, "vram_gb": 20.709735} {"adam_lr": 9.713207455460894e-05, "dt_ms": 165.344715, "global_tokens_per_sec": 99089.952698, "grad_norm": 40.5, "loss": 3.883976, "lr_stage": "cosine_decay", "mfu": 56.921088, "stage_name": "permutation_cycle", "status": "training", "step": 395, "tokens_per_sec_per_gpu": 99089.952698, "vram_gb": 20.709735} {"adam_lr": 9.710287263936484e-05, "dt_ms": 162.500143, "global_tokens_per_sec": 100824.526627, "grad_norm": 41.75, "loss": 3.017439, "lr_stage": "cosine_decay", "mfu": 57.917494, "stage_name": "permutation_cycle", "status": "training", "step": 396, "tokens_per_sec_per_gpu": 100824.526627, "vram_gb": 20.709735} {"adam_lr": 9.70735272406019e-05, "dt_ms": 165.192127, "global_tokens_per_sec": 99181.482041, "grad_norm": 35.0, "loss": 3.450592, "lr_stage": "cosine_decay", "mfu": 56.973666, "stage_name": "permutation_cycle", "status": "training", "step": 397, "tokens_per_sec_per_gpu": 99181.482041, "vram_gb": 20.709735} {"adam_lr": 9.704403844771128e-05, "dt_ms": 161.996126, "global_tokens_per_sec": 101138.220937, "grad_norm": 38.75, "loss": 3.573406, "lr_stage": "cosine_decay", "mfu": 58.097692, "stage_name": "permutation_cycle", "status": "training", "step": 398, "tokens_per_sec_per_gpu": 101138.220937, "vram_gb": 20.709735} {"adam_lr": 9.701440635052094e-05, "dt_ms": 163.324118, "global_tokens_per_sec": 100315.864152, "grad_norm": 41.0, "loss": 3.20727, "lr_stage": "cosine_decay", "mfu": 57.625299, "stage_name": "permutation_cycle", "status": "training", "step": 399, "tokens_per_sec_per_gpu": 100315.864152, "vram_gb": 20.709735} {"adam_lr": 9.698463103929542e-05, "dt_ms": 162.864923, "global_tokens_per_sec": 100598.702595, "grad_norm": 33.25, "loss": 3.379593, "lr_stage": "cosine_decay", "mfu": 57.787772, "stage_name": "permutation_cycle", "status": "training", "step": 400, "tokens_per_sec_per_gpu": 100598.702595, "vram_gb": 20.709735} {"adam_lr": 9.695471260473545e-05, "dt_ms": 165.035248, "global_tokens_per_sec": 99275.762106, "grad_norm": 51.5, "loss": 5.156081, "lr_stage": "cosine_decay", "mfu": 57.027824, "stage_name": "permutation_cycle", "status": "training", "step": 401, "tokens_per_sec_per_gpu": 99275.762106, "vram_gb": 20.709735} {"adam_lr": 9.69246511379778e-05, "dt_ms": 162.399769, "global_tokens_per_sec": 100886.843116, "grad_norm": 32.75, "loss": 3.603152, "lr_stage": "cosine_decay", "mfu": 57.953291, "stage_name": "permutation_cycle", "status": "training", "step": 402, "tokens_per_sec_per_gpu": 100886.843116, "vram_gb": 20.709735} {"adam_lr": 9.68944467305949e-05, "dt_ms": 162.107468, "global_tokens_per_sec": 101068.755421, "grad_norm": 56.75, "loss": 4.783846, "lr_stage": "cosine_decay", "mfu": 58.057789, "stage_name": "permutation_cycle", "status": "training", "step": 403, "tokens_per_sec_per_gpu": 101068.755421, "vram_gb": 20.709735} {"adam_lr": 9.686409947459458e-05, "dt_ms": 164.393187, "global_tokens_per_sec": 99663.497873, "grad_norm": 73.0, "loss": 4.823124, "lr_stage": "cosine_decay", "mfu": 57.250554, "stage_name": "permutation_cycle", "status": "training", "step": 404, "tokens_per_sec_per_gpu": 99663.497873, "vram_gb": 20.709735} {"adam_lr": 9.683360946241989e-05, "dt_ms": 163.651705, "global_tokens_per_sec": 100115.058509, "grad_norm": 28.625, "loss": 4.091721, "lr_stage": "cosine_decay", "mfu": 57.509948, "stage_name": "permutation_cycle", "status": "training", "step": 405, "tokens_per_sec_per_gpu": 100115.058509, "vram_gb": 20.709735} {"adam_lr": 9.680297678694867e-05, "dt_ms": 162.161589, "global_tokens_per_sec": 101035.023981, "grad_norm": 33.5, "loss": 3.892033, "lr_stage": "cosine_decay", "mfu": 58.038412, "stage_name": "permutation_cycle", "status": "training", "step": 406, "tokens_per_sec_per_gpu": 101035.023981, "vram_gb": 20.709735} {"adam_lr": 9.677220154149337e-05, "dt_ms": 163.024664, "global_tokens_per_sec": 100500.130505, "grad_norm": 48.5, "loss": 4.576951, "lr_stage": "cosine_decay", "mfu": 57.731148, "stage_name": "permutation_cycle", "status": "training", "step": 407, "tokens_per_sec_per_gpu": 100500.130505, "vram_gb": 20.709735} {"adam_lr": 9.674128381980072e-05, "dt_ms": 164.144993, "global_tokens_per_sec": 99814.193036, "grad_norm": 37.25, "loss": 3.731562, "lr_stage": "cosine_decay", "mfu": 57.33712, "stage_name": "permutation_cycle", "status": "training", "step": 408, "tokens_per_sec_per_gpu": 99814.193036, "vram_gb": 20.709735} {"adam_lr": 9.671022371605148e-05, "dt_ms": 163.568258, "global_tokens_per_sec": 100166.133526, "grad_norm": 36.0, "loss": 3.326595, "lr_stage": "cosine_decay", "mfu": 57.539288, "stage_name": "permutation_cycle", "status": "training", "step": 409, "tokens_per_sec_per_gpu": 100166.133526, "vram_gb": 20.709735} {"adam_lr": 9.667902132486009e-05, "dt_ms": 162.497759, "global_tokens_per_sec": 100826.005936, "grad_norm": 37.0, "loss": 3.789153, "lr_stage": "cosine_decay", "mfu": 57.918344, "stage_name": "permutation_cycle", "status": "training", "step": 410, "tokens_per_sec_per_gpu": 100826.005936, "vram_gb": 20.709735} {"adam_lr": 9.664767674127447e-05, "dt_ms": 163.341284, "global_tokens_per_sec": 100305.321588, "grad_norm": 36.5, "loss": 3.605083, "lr_stage": "cosine_decay", "mfu": 57.619243, "stage_name": "permutation_cycle", "status": "training", "step": 411, "tokens_per_sec_per_gpu": 100305.321588, "vram_gb": 20.709735} {"adam_lr": 9.661619006077561e-05, "dt_ms": 163.283825, "global_tokens_per_sec": 100340.618601, "grad_norm": 33.75, "loss": 3.429042, "lr_stage": "cosine_decay", "mfu": 57.639519, "stage_name": "permutation_cycle", "status": "training", "step": 412, "tokens_per_sec_per_gpu": 100340.618601, "vram_gb": 20.709735} {"adam_lr": 9.658456137927745e-05, "dt_ms": 163.682461, "global_tokens_per_sec": 100096.246852, "grad_norm": 40.5, "loss": 3.29184, "lr_stage": "cosine_decay", "mfu": 57.499142, "stage_name": "permutation_cycle", "status": "training", "step": 413, "tokens_per_sec_per_gpu": 100096.246852, "vram_gb": 20.709735} {"adam_lr": 9.655279079312642e-05, "dt_ms": 162.933826, "global_tokens_per_sec": 100556.160481, "grad_norm": 40.0, "loss": 3.425653, "lr_stage": "cosine_decay", "mfu": 57.763334, "stage_name": "permutation_cycle", "status": "training", "step": 414, "tokens_per_sec_per_gpu": 100556.160481, "vram_gb": 20.709735} {"adam_lr": 9.652087839910124e-05, "dt_ms": 162.937164, "global_tokens_per_sec": 100554.100531, "grad_norm": 38.25, "loss": 2.851384, "lr_stage": "cosine_decay", "mfu": 57.762151, "stage_name": "permutation_cycle", "status": "training", "step": 415, "tokens_per_sec_per_gpu": 100554.100531, "vram_gb": 20.709735} {"adam_lr": 9.648882429441257e-05, "dt_ms": 162.732601, "global_tokens_per_sec": 100680.502141, "grad_norm": 31.75, "loss": 3.470522, "lr_stage": "cosine_decay", "mfu": 57.834761, "stage_name": "permutation_cycle", "status": "training", "step": 416, "tokens_per_sec_per_gpu": 100680.502141, "vram_gb": 20.709735} {"adam_lr": 9.645662857670281e-05, "dt_ms": 163.734674, "global_tokens_per_sec": 100064.326965, "grad_norm": 44.25, "loss": 3.203523, "lr_stage": "cosine_decay", "mfu": 57.480806, "stage_name": "permutation_cycle", "status": "training", "step": 417, "tokens_per_sec_per_gpu": 100064.326965, "vram_gb": 20.709735} {"adam_lr": 9.642429134404569e-05, "dt_ms": 162.971973, "global_tokens_per_sec": 100532.623225, "grad_norm": 31.75, "loss": 3.305053, "lr_stage": "cosine_decay", "mfu": 57.749814, "stage_name": "permutation_cycle", "status": "training", "step": 418, "tokens_per_sec_per_gpu": 100532.623225, "vram_gb": 20.709735} {"adam_lr": 9.6391812694946e-05, "dt_ms": 161.772013, "global_tokens_per_sec": 101278.334401, "grad_norm": 36.25, "loss": 3.923981, "lr_stage": "cosine_decay", "mfu": 58.178179, "stage_name": "permutation_cycle", "status": "training", "step": 419, "tokens_per_sec_per_gpu": 101278.334401, "vram_gb": 20.709735} {"adam_lr": 9.635919272833938e-05, "dt_ms": 163.184643, "global_tokens_per_sec": 100401.604708, "grad_norm": 49.5, "loss": 3.449165, "lr_stage": "cosine_decay", "mfu": 57.674551, "stage_name": "permutation_cycle", "status": "training", "step": 420, "tokens_per_sec_per_gpu": 100401.604708, "vram_gb": 20.709735} {"adam_lr": 9.632643154359186e-05, "dt_ms": 162.180901, "global_tokens_per_sec": 101022.993102, "grad_norm": 36.25, "loss": 3.273369, "lr_stage": "cosine_decay", "mfu": 58.031501, "stage_name": "permutation_cycle", "status": "training", "step": 421, "tokens_per_sec_per_gpu": 101022.993102, "vram_gb": 20.709735} {"adam_lr": 9.629352924049975e-05, "dt_ms": 162.976503, "global_tokens_per_sec": 100529.828908, "grad_norm": 36.0, "loss": 2.417221, "lr_stage": "cosine_decay", "mfu": 57.748208, "stage_name": "permutation_cycle", "status": "training", "step": 422, "tokens_per_sec_per_gpu": 100529.828908, "vram_gb": 20.709735} {"adam_lr": 9.626048591928912e-05, "dt_ms": 162.910461, "global_tokens_per_sec": 100570.582494, "grad_norm": 37.0, "loss": 3.226226, "lr_stage": "cosine_decay", "mfu": 57.771619, "stage_name": "permutation_cycle", "status": "training", "step": 423, "tokens_per_sec_per_gpu": 100570.582494, "vram_gb": 20.709735} {"adam_lr": 9.622730168061567e-05, "dt_ms": 164.419651, "global_tokens_per_sec": 99647.456355, "grad_norm": 45.75, "loss": 4.023022, "lr_stage": "cosine_decay", "mfu": 57.24134, "stage_name": "permutation_cycle", "status": "training", "step": 424, "tokens_per_sec_per_gpu": 99647.456355, "vram_gb": 20.709735} {"adam_lr": 9.619397662556435e-05, "dt_ms": 162.595034, "global_tokens_per_sec": 100765.685351, "grad_norm": 38.5, "loss": 3.476884, "lr_stage": "cosine_decay", "mfu": 57.883693, "stage_name": "permutation_cycle", "status": "training", "step": 425, "tokens_per_sec_per_gpu": 100765.685351, "vram_gb": 20.709735} {"adam_lr": 9.616051085564906e-05, "dt_ms": 164.119482, "global_tokens_per_sec": 99829.708188, "grad_norm": 35.75, "loss": 3.463539, "lr_stage": "cosine_decay", "mfu": 57.346032, "stage_name": "permutation_cycle", "status": "training", "step": 426, "tokens_per_sec_per_gpu": 99829.708188, "vram_gb": 20.709735} {"adam_lr": 9.612690447281231e-05, "dt_ms": 162.524939, "global_tokens_per_sec": 100809.144386, "grad_norm": 30.625, "loss": 2.812146, "lr_stage": "cosine_decay", "mfu": 57.908658, "stage_name": "permutation_cycle", "status": "training", "step": 427, "tokens_per_sec_per_gpu": 100809.144386, "vram_gb": 20.709735} {"adam_lr": 9.609315757942503e-05, "dt_ms": 162.037373, "global_tokens_per_sec": 101112.476327, "grad_norm": 37.5, "loss": 4.179177, "lr_stage": "cosine_decay", "mfu": 58.082904, "stage_name": "permutation_cycle", "status": "training", "step": 428, "tokens_per_sec_per_gpu": 101112.476327, "vram_gb": 20.709735} {"adam_lr": 9.605927027828608e-05, "dt_ms": 162.964344, "global_tokens_per_sec": 100537.329795, "grad_norm": 34.5, "loss": 2.820387, "lr_stage": "cosine_decay", "mfu": 57.752517, "stage_name": "permutation_cycle", "status": "training", "step": 429, "tokens_per_sec_per_gpu": 100537.329795, "vram_gb": 20.709735} {"adam_lr": 9.602524267262203e-05, "dt_ms": 162.689924, "global_tokens_per_sec": 100706.912715, "grad_norm": 24.875, "loss": 3.151185, "lr_stage": "cosine_decay", "mfu": 57.849932, "stage_name": "permutation_cycle", "status": "training", "step": 430, "tokens_per_sec_per_gpu": 100706.912715, "vram_gb": 20.709735} {"adam_lr": 9.599107486608689e-05, "dt_ms": 161.716461, "global_tokens_per_sec": 101313.124714, "grad_norm": 40.0, "loss": 3.378897, "lr_stage": "cosine_decay", "mfu": 58.198164, "stage_name": "permutation_cycle", "status": "training", "step": 431, "tokens_per_sec_per_gpu": 101313.124714, "vram_gb": 20.709735} {"adam_lr": 9.595676696276172e-05, "dt_ms": 162.577391, "global_tokens_per_sec": 100776.620491, "grad_norm": 46.75, "loss": 3.898153, "lr_stage": "cosine_decay", "mfu": 57.889975, "stage_name": "permutation_cycle", "status": "training", "step": 432, "tokens_per_sec_per_gpu": 100776.620491, "vram_gb": 20.709735} {"adam_lr": 9.592231906715436e-05, "dt_ms": 163.326979, "global_tokens_per_sec": 100314.106904, "grad_norm": 49.0, "loss": 3.742518, "lr_stage": "cosine_decay", "mfu": 57.624289, "stage_name": "permutation_cycle", "status": "training", "step": 433, "tokens_per_sec_per_gpu": 100314.106904, "vram_gb": 20.709735} {"adam_lr": 9.588773128419906e-05, "dt_ms": 162.535429, "global_tokens_per_sec": 100802.637928, "grad_norm": 39.75, "loss": 3.441218, "lr_stage": "cosine_decay", "mfu": 57.90492, "stage_name": "permutation_cycle", "status": "training", "step": 434, "tokens_per_sec_per_gpu": 100802.637928, "vram_gb": 20.709735} {"adam_lr": 9.58530037192562e-05, "dt_ms": 162.324429, "global_tokens_per_sec": 100933.668121, "grad_norm": 53.5, "loss": 4.43519, "lr_stage": "cosine_decay", "mfu": 57.980189, "stage_name": "permutation_cycle", "status": "training", "step": 435, "tokens_per_sec_per_gpu": 100933.668121, "vram_gb": 20.709735} {"adam_lr": 9.581813647811198e-05, "dt_ms": 164.536476, "global_tokens_per_sec": 99576.704114, "grad_norm": 39.0, "loss": 3.717991, "lr_stage": "cosine_decay", "mfu": 57.200697, "stage_name": "permutation_cycle", "status": "training", "step": 436, "tokens_per_sec_per_gpu": 99576.704114, "vram_gb": 20.709735} {"adam_lr": 9.578312966697807e-05, "dt_ms": 162.093163, "global_tokens_per_sec": 101077.674984, "grad_norm": 49.5, "loss": 3.726934, "lr_stage": "cosine_decay", "mfu": 58.062912, "stage_name": "permutation_cycle", "status": "training", "step": 437, "tokens_per_sec_per_gpu": 101077.674984, "vram_gb": 20.709735} {"adam_lr": 9.574798339249125e-05, "dt_ms": 163.138151, "global_tokens_per_sec": 100430.217473, "grad_norm": 43.5, "loss": 3.543118, "lr_stage": "cosine_decay", "mfu": 57.690988, "stage_name": "permutation_cycle", "status": "training", "step": 438, "tokens_per_sec_per_gpu": 100430.217473, "vram_gb": 20.709735} {"adam_lr": 9.571269776171319e-05, "dt_ms": 163.572311, "global_tokens_per_sec": 100163.651535, "grad_norm": 44.25, "loss": 4.374624, "lr_stage": "cosine_decay", "mfu": 57.537862, "stage_name": "permutation_cycle", "status": "training", "step": 439, "tokens_per_sec_per_gpu": 100163.651535, "vram_gb": 20.709735} {"adam_lr": 9.567727288213005e-05, "dt_ms": 163.970947, "global_tokens_per_sec": 99920.139959, "grad_norm": 49.75, "loss": 3.44439, "lr_stage": "cosine_decay", "mfu": 57.39798, "stage_name": "permutation_cycle", "status": "training", "step": 440, "tokens_per_sec_per_gpu": 99920.139959, "vram_gb": 20.709735} {"adam_lr": 9.564170886165214e-05, "dt_ms": 164.101839, "global_tokens_per_sec": 99840.441114, "grad_norm": 35.25, "loss": 3.09346, "lr_stage": "cosine_decay", "mfu": 57.352197, "stage_name": "permutation_cycle", "status": "training", "step": 441, "tokens_per_sec_per_gpu": 99840.441114, "vram_gb": 20.709735} {"adam_lr": 9.560600580861365e-05, "dt_ms": 161.63373, "global_tokens_per_sec": 101364.981224, "grad_norm": 36.25, "loss": 2.819063, "lr_stage": "cosine_decay", "mfu": 58.227952, "stage_name": "permutation_cycle", "status": "training", "step": 442, "tokens_per_sec_per_gpu": 101364.981224, "vram_gb": 20.709735} {"adam_lr": 9.557016383177227e-05, "dt_ms": 162.648678, "global_tokens_per_sec": 100732.451189, "grad_norm": 42.0, "loss": 5.128052, "lr_stage": "cosine_decay", "mfu": 57.864602, "stage_name": "permutation_cycle", "status": "training", "step": 443, "tokens_per_sec_per_gpu": 100732.451189, "vram_gb": 20.709735} {"adam_lr": 9.553418304030886e-05, "dt_ms": 163.547277, "global_tokens_per_sec": 100178.983444, "grad_norm": 52.5, "loss": 3.754719, "lr_stage": "cosine_decay", "mfu": 57.546669, "stage_name": "permutation_cycle", "status": "training", "step": 444, "tokens_per_sec_per_gpu": 100178.983444, "vram_gb": 20.709735} {"adam_lr": 9.549806354382717e-05, "dt_ms": 162.928343, "global_tokens_per_sec": 100559.544868, "grad_norm": 32.75, "loss": 3.590368, "lr_stage": "cosine_decay", "mfu": 57.765278, "stage_name": "permutation_cycle", "status": "training", "step": 445, "tokens_per_sec_per_gpu": 100559.544868, "vram_gb": 20.709735} {"adam_lr": 9.546180545235344e-05, "dt_ms": 161.901474, "global_tokens_per_sec": 101197.349198, "grad_norm": 35.0, "loss": 4.206911, "lr_stage": "cosine_decay", "mfu": 58.131658, "stage_name": "permutation_cycle", "status": "training", "step": 446, "tokens_per_sec_per_gpu": 101197.349198, "vram_gb": 20.709735} {"adam_lr": 9.54254088763361e-05, "dt_ms": 164.214373, "global_tokens_per_sec": 99772.022005, "grad_norm": 37.0, "loss": 2.382088, "lr_stage": "cosine_decay", "mfu": 57.312895, "stage_name": "permutation_cycle", "status": "training", "step": 447, "tokens_per_sec_per_gpu": 99772.022005, "vram_gb": 20.709735} {"adam_lr": 9.538887392664544e-05, "dt_ms": 161.487341, "global_tokens_per_sec": 101456.86904, "grad_norm": 40.0, "loss": 3.0303, "lr_stage": "cosine_decay", "mfu": 58.280736, "stage_name": "permutation_cycle", "status": "training", "step": 448, "tokens_per_sec_per_gpu": 101456.86904, "vram_gb": 20.709735} {"adam_lr": 9.535220071457325e-05, "dt_ms": 163.561821, "global_tokens_per_sec": 100170.075764, "grad_norm": 33.25, "loss": 3.675254, "lr_stage": "cosine_decay", "mfu": 57.541552, "stage_name": "permutation_cycle", "status": "training", "step": 449, "tokens_per_sec_per_gpu": 100170.075764, "vram_gb": 20.709735} {"adam_lr": 9.53153893518325e-05, "dt_ms": 164.348602, "global_tokens_per_sec": 99690.534457, "grad_norm": 32.0, "loss": 3.985155, "lr_stage": "cosine_decay", "mfu": 57.266085, "stage_name": "permutation_cycle", "status": "training", "step": 450, "tokens_per_sec_per_gpu": 99690.534457, "vram_gb": 20.709735} {"adam_lr": 9.527843995055699e-05, "dt_ms": 163.491249, "global_tokens_per_sec": 100213.314729, "grad_norm": 51.0, "loss": 3.395137, "lr_stage": "cosine_decay", "mfu": 57.56639, "stage_name": "permutation_cycle", "status": "training", "step": 451, "tokens_per_sec_per_gpu": 100213.314729, "vram_gb": 20.709735} {"adam_lr": 9.524135262330098e-05, "dt_ms": 162.092447, "global_tokens_per_sec": 101078.121003, "grad_norm": 39.25, "loss": 3.132516, "lr_stage": "cosine_decay", "mfu": 58.063168, "stage_name": "permutation_cycle", "status": "training", "step": 452, "tokens_per_sec_per_gpu": 101078.121003, "vram_gb": 20.709735} {"adam_lr": 9.520412748303893e-05, "dt_ms": 162.708044, "global_tokens_per_sec": 100695.697594, "grad_norm": 42.0, "loss": 3.344036, "lr_stage": "cosine_decay", "mfu": 57.84349, "stage_name": "permutation_cycle", "status": "training", "step": 453, "tokens_per_sec_per_gpu": 100695.697594, "vram_gb": 20.709735} {"adam_lr": 9.516676464316505e-05, "dt_ms": 163.769245, "global_tokens_per_sec": 100043.203992, "grad_norm": 37.25, "loss": 3.189726, "lr_stage": "cosine_decay", "mfu": 57.468672, "stage_name": "permutation_cycle", "status": "training", "step": 454, "tokens_per_sec_per_gpu": 100043.203992, "vram_gb": 20.709735} {"adam_lr": 9.512926421749304e-05, "dt_ms": 164.852142, "global_tokens_per_sec": 99386.030221, "grad_norm": 32.75, "loss": 4.425182, "lr_stage": "cosine_decay", "mfu": 57.091166, "stage_name": "permutation_cycle", "status": "training", "step": 455, "tokens_per_sec_per_gpu": 99386.030221, "vram_gb": 20.709735} {"adam_lr": 9.50916263202557e-05, "dt_ms": 163.953543, "global_tokens_per_sec": 99930.747023, "grad_norm": 38.75, "loss": 4.04204, "lr_stage": "cosine_decay", "mfu": 57.404073, "stage_name": "permutation_cycle", "status": "training", "step": 456, "tokens_per_sec_per_gpu": 99930.747023, "vram_gb": 20.709735} {"adam_lr": 9.505385106610458e-05, "dt_ms": 162.808895, "global_tokens_per_sec": 100633.322208, "grad_norm": 40.0, "loss": 4.55525, "lr_stage": "cosine_decay", "mfu": 57.807659, "stage_name": "permutation_cycle", "status": "training", "step": 457, "tokens_per_sec_per_gpu": 100633.322208, "vram_gb": 20.709735} {"adam_lr": 9.501593857010969e-05, "dt_ms": 163.959742, "global_tokens_per_sec": 99926.968906, "grad_norm": 44.5, "loss": 2.897025, "lr_stage": "cosine_decay", "mfu": 57.401902, "stage_name": "permutation_cycle", "status": "training", "step": 458, "tokens_per_sec_per_gpu": 99926.968906, "vram_gb": 20.709735} {"adam_lr": 9.497788894775901e-05, "dt_ms": 163.468599, "global_tokens_per_sec": 100227.200014, "grad_norm": 37.5, "loss": 2.627605, "lr_stage": "cosine_decay", "mfu": 57.574367, "stage_name": "permutation_cycle", "status": "training", "step": 459, "tokens_per_sec_per_gpu": 100227.200014, "vram_gb": 20.709735} {"adam_lr": 9.493970231495835e-05, "dt_ms": 162.759781, "global_tokens_per_sec": 100663.689218, "grad_norm": 31.75, "loss": 3.837421, "lr_stage": "cosine_decay", "mfu": 57.825103, "stage_name": "permutation_cycle", "status": "training", "step": 460, "tokens_per_sec_per_gpu": 100663.689218, "vram_gb": 20.709735} {"adam_lr": 9.490137878803079e-05, "dt_ms": 164.927006, "global_tokens_per_sec": 99340.917054, "grad_norm": 38.5, "loss": 2.810103, "lr_stage": "cosine_decay", "mfu": 57.065252, "stage_name": "permutation_cycle", "status": "training", "step": 461, "tokens_per_sec_per_gpu": 99340.917054, "vram_gb": 20.709735} {"adam_lr": 9.486291848371643e-05, "dt_ms": 164.031744, "global_tokens_per_sec": 99883.105551, "grad_norm": 35.75, "loss": 3.233775, "lr_stage": "cosine_decay", "mfu": 57.376706, "stage_name": "permutation_cycle", "status": "training", "step": 462, "tokens_per_sec_per_gpu": 99883.105551, "vram_gb": 20.709735} {"adam_lr": 9.482432151917203e-05, "dt_ms": 161.830902, "global_tokens_per_sec": 101241.479763, "grad_norm": 36.25, "loss": 4.424249, "lr_stage": "cosine_decay", "mfu": 58.157008, "stage_name": "permutation_cycle", "status": "training", "step": 463, "tokens_per_sec_per_gpu": 101241.479763, "vram_gb": 20.709735} {"adam_lr": 9.478558801197065e-05, "dt_ms": 163.244247, "global_tokens_per_sec": 100364.945517, "grad_norm": 34.5, "loss": 3.755647, "lr_stage": "cosine_decay", "mfu": 57.653493, "stage_name": "permutation_cycle", "status": "training", "step": 464, "tokens_per_sec_per_gpu": 100364.945517, "vram_gb": 20.709735} {"adam_lr": 9.474671808010126e-05, "dt_ms": 161.39555, "global_tokens_per_sec": 101514.571021, "grad_norm": 33.75, "loss": 3.657066, "lr_stage": "cosine_decay", "mfu": 58.313882, "stage_name": "permutation_cycle", "status": "training", "step": 465, "tokens_per_sec_per_gpu": 101514.571021, "vram_gb": 20.709735} {"adam_lr": 9.47077118419684e-05, "dt_ms": 162.126541, "global_tokens_per_sec": 101056.865119, "grad_norm": 32.5, "loss": 2.947904, "lr_stage": "cosine_decay", "mfu": 58.050958, "stage_name": "permutation_cycle", "status": "training", "step": 466, "tokens_per_sec_per_gpu": 101056.865119, "vram_gb": 20.709735} {"adam_lr": 9.466856941639188e-05, "dt_ms": 161.642075, "global_tokens_per_sec": 101359.748333, "grad_norm": 25.75, "loss": 3.949063, "lr_stage": "cosine_decay", "mfu": 58.224946, "stage_name": "permutation_cycle", "status": "training", "step": 467, "tokens_per_sec_per_gpu": 101359.748333, "vram_gb": 20.709735} {"adam_lr": 9.462929092260628e-05, "dt_ms": 162.243843, "global_tokens_per_sec": 100983.801229, "grad_norm": 35.75, "loss": 3.417803, "lr_stage": "cosine_decay", "mfu": 58.008988, "stage_name": "permutation_cycle", "status": "training", "step": 468, "tokens_per_sec_per_gpu": 100983.801229, "vram_gb": 20.709735} {"adam_lr": 9.458987648026071e-05, "dt_ms": 163.750172, "global_tokens_per_sec": 100054.856943, "grad_norm": 31.25, "loss": 2.934762, "lr_stage": "cosine_decay", "mfu": 57.475366, "stage_name": "permutation_cycle", "status": "training", "step": 469, "tokens_per_sec_per_gpu": 100054.856943, "vram_gb": 20.709735} {"adam_lr": 9.45503262094184e-05, "dt_ms": 164.345264, "global_tokens_per_sec": 99692.559176, "grad_norm": 29.5, "loss": 2.888133, "lr_stage": "cosine_decay", "mfu": 57.267248, "stage_name": "permutation_cycle", "status": "training", "step": 470, "tokens_per_sec_per_gpu": 99692.559176, "vram_gb": 20.709735} {"adam_lr": 9.451064023055633e-05, "dt_ms": 162.200689, "global_tokens_per_sec": 101010.668137, "grad_norm": 40.75, "loss": 3.190732, "lr_stage": "cosine_decay", "mfu": 58.024421, "stage_name": "permutation_cycle", "status": "training", "step": 471, "tokens_per_sec_per_gpu": 101010.668137, "vram_gb": 20.709735} {"adam_lr": 9.447081866456489e-05, "dt_ms": 163.665533, "global_tokens_per_sec": 100106.59968, "grad_norm": 54.0, "loss": 3.791946, "lr_stage": "cosine_decay", "mfu": 57.505089, "stage_name": "permutation_cycle", "status": "training", "step": 472, "tokens_per_sec_per_gpu": 100106.59968, "vram_gb": 20.709735} {"adam_lr": 9.443086163274745e-05, "dt_ms": 165.617228, "global_tokens_per_sec": 98926.90659, "grad_norm": 29.5, "loss": 3.575294, "lr_stage": "cosine_decay", "mfu": 56.827428, "stage_name": "permutation_cycle", "status": "training", "step": 473, "tokens_per_sec_per_gpu": 98926.90659, "vram_gb": 20.709735} {"adam_lr": 9.439076925682006e-05, "dt_ms": 163.054228, "global_tokens_per_sec": 100481.908492, "grad_norm": 56.0, "loss": 4.227719, "lr_stage": "cosine_decay", "mfu": 57.720681, "stage_name": "permutation_cycle", "status": "training", "step": 474, "tokens_per_sec_per_gpu": 100481.908492, "vram_gb": 20.709735} {"adam_lr": 9.435054165891109e-05, "dt_ms": 164.420843, "global_tokens_per_sec": 99646.733885, "grad_norm": 24.0, "loss": 3.492862, "lr_stage": "cosine_decay", "mfu": 57.240925, "stage_name": "permutation_cycle", "status": "training", "step": 475, "tokens_per_sec_per_gpu": 99646.733885, "vram_gb": 20.709735} {"adam_lr": 9.431017896156074e-05, "dt_ms": 164.42585, "global_tokens_per_sec": 99643.699628, "grad_norm": 32.5, "loss": 3.962362, "lr_stage": "cosine_decay", "mfu": 57.239182, "stage_name": "permutation_cycle", "status": "training", "step": 476, "tokens_per_sec_per_gpu": 99643.699628, "vram_gb": 20.709735} {"adam_lr": 9.42696812877208e-05, "dt_ms": 163.594007, "global_tokens_per_sec": 100150.367676, "grad_norm": 50.0, "loss": 3.791034, "lr_stage": "cosine_decay", "mfu": 57.530231, "stage_name": "permutation_cycle", "status": "training", "step": 477, "tokens_per_sec_per_gpu": 100150.367676, "vram_gb": 20.709735} {"adam_lr": 9.42290487607542e-05, "dt_ms": 163.689375, "global_tokens_per_sec": 100092.018848, "grad_norm": 40.25, "loss": 4.703944, "lr_stage": "cosine_decay", "mfu": 57.496713, "stage_name": "permutation_cycle", "status": "training", "step": 478, "tokens_per_sec_per_gpu": 100092.018848, "vram_gb": 20.709735} {"adam_lr": 9.418828150443469e-05, "dt_ms": 165.106535, "global_tokens_per_sec": 99232.898348, "grad_norm": 35.0, "loss": 3.240937, "lr_stage": "cosine_decay", "mfu": 57.003201, "stage_name": "permutation_cycle", "status": "training", "step": 479, "tokens_per_sec_per_gpu": 99232.898348, "vram_gb": 20.709735} {"adam_lr": 9.414737964294636e-05, "dt_ms": 163.865566, "global_tokens_per_sec": 99984.398032, "grad_norm": 35.5, "loss": 3.332761, "lr_stage": "cosine_decay", "mfu": 57.434892, "stage_name": "permutation_cycle", "status": "training", "step": 480, "tokens_per_sec_per_gpu": 99984.398032, "vram_gb": 20.709735} {"adam_lr": 9.410634330088339e-05, "dt_ms": 162.193298, "global_tokens_per_sec": 101015.271085, "grad_norm": 69.0, "loss": 4.226597, "lr_stage": "cosine_decay", "mfu": 58.027065, "stage_name": "permutation_cycle", "status": "training", "step": 481, "tokens_per_sec_per_gpu": 101015.271085, "vram_gb": 20.709735} {"adam_lr": 9.40651726032496e-05, "dt_ms": 162.730217, "global_tokens_per_sec": 100681.977226, "grad_norm": 41.0, "loss": 3.826239, "lr_stage": "cosine_decay", "mfu": 57.835608, "stage_name": "permutation_cycle", "status": "training", "step": 482, "tokens_per_sec_per_gpu": 100681.977226, "vram_gb": 20.709735} {"adam_lr": 9.40238676754581e-05, "dt_ms": 162.646055, "global_tokens_per_sec": 100734.07546, "grad_norm": 28.125, "loss": 2.159067, "lr_stage": "cosine_decay", "mfu": 57.865535, "stage_name": "permutation_cycle", "status": "training", "step": 483, "tokens_per_sec_per_gpu": 100734.07546, "vram_gb": 20.709735} {"adam_lr": 9.398242864333083e-05, "dt_ms": 163.102388, "global_tokens_per_sec": 100452.238392, "grad_norm": 43.25, "loss": 3.514446, "lr_stage": "cosine_decay", "mfu": 57.703637, "stage_name": "permutation_cycle", "status": "training", "step": 484, "tokens_per_sec_per_gpu": 100452.238392, "vram_gb": 20.709735} {"adam_lr": 9.394085563309827e-05, "dt_ms": 163.644552, "global_tokens_per_sec": 100119.434327, "grad_norm": 35.5, "loss": 2.929812, "lr_stage": "cosine_decay", "mfu": 57.512462, "stage_name": "permutation_cycle", "status": "training", "step": 485, "tokens_per_sec_per_gpu": 100119.434327, "vram_gb": 20.709735} {"adam_lr": 9.389914877139903e-05, "dt_ms": 163.698912, "global_tokens_per_sec": 100086.187704, "grad_norm": 36.25, "loss": 2.920539, "lr_stage": "cosine_decay", "mfu": 57.493364, "stage_name": "permutation_cycle", "status": "training", "step": 486, "tokens_per_sec_per_gpu": 100086.187704, "vram_gb": 20.709735} {"adam_lr": 9.385730818527944e-05, "dt_ms": 161.959648, "global_tokens_per_sec": 101161.000218, "grad_norm": 51.5, "loss": 3.974103, "lr_stage": "cosine_decay", "mfu": 58.110778, "stage_name": "permutation_cycle", "status": "training", "step": 487, "tokens_per_sec_per_gpu": 101161.000218, "vram_gb": 20.709735} {"adam_lr": 9.381533400219318e-05, "dt_ms": 162.039995, "global_tokens_per_sec": 101110.83983, "grad_norm": 30.875, "loss": 2.897365, "lr_stage": "cosine_decay", "mfu": 58.081963, "stage_name": "permutation_cycle", "status": "training", "step": 488, "tokens_per_sec_per_gpu": 101110.83983, "vram_gb": 20.709735} {"adam_lr": 9.37732263500009e-05, "dt_ms": 159.805059, "global_tokens_per_sec": 102524.914156, "grad_norm": 45.25, "loss": 3.858263, "lr_stage": "cosine_decay", "mfu": 58.894262, "stage_name": "permutation_cycle", "status": "training", "step": 489, "tokens_per_sec_per_gpu": 102524.914156, "vram_gb": 20.709735} {"adam_lr": 9.373098535696979e-05, "dt_ms": 162.898064, "global_tokens_per_sec": 100578.236671, "grad_norm": 35.75, "loss": 2.977796, "lr_stage": "cosine_decay", "mfu": 57.776016, "stage_name": "permutation_cycle", "status": "training", "step": 490, "tokens_per_sec_per_gpu": 100578.236671, "vram_gb": 20.709735} {"adam_lr": 9.368861115177327e-05, "dt_ms": 162.749529, "global_tokens_per_sec": 100670.030274, "grad_norm": 43.0, "loss": 3.364029, "lr_stage": "cosine_decay", "mfu": 57.828745, "stage_name": "permutation_cycle", "status": "training", "step": 491, "tokens_per_sec_per_gpu": 100670.030274, "vram_gb": 20.709735} {"adam_lr": 9.364610386349049e-05, "dt_ms": 162.089109, "global_tokens_per_sec": 101080.20248, "grad_norm": 62.0, "loss": 3.319457, "lr_stage": "cosine_decay", "mfu": 58.064364, "stage_name": "permutation_cycle", "status": "training", "step": 492, "tokens_per_sec_per_gpu": 101080.20248, "vram_gb": 20.709735} {"adam_lr": 9.360346362160604e-05, "dt_ms": 162.697792, "global_tokens_per_sec": 100702.042684, "grad_norm": 37.0, "loss": 3.605347, "lr_stage": "cosine_decay", "mfu": 57.847135, "stage_name": "permutation_cycle", "status": "training", "step": 493, "tokens_per_sec_per_gpu": 100702.042684, "vram_gb": 20.709735} {"adam_lr": 9.356069055600948e-05, "dt_ms": 163.763523, "global_tokens_per_sec": 100046.699593, "grad_norm": 29.625, "loss": 3.587254, "lr_stage": "cosine_decay", "mfu": 57.47068, "stage_name": "permutation_cycle", "status": "training", "step": 494, "tokens_per_sec_per_gpu": 100046.699593, "vram_gb": 20.709735} {"adam_lr": 9.351778479699499e-05, "dt_ms": 161.959648, "global_tokens_per_sec": 101161.000218, "grad_norm": 47.0, "loss": 3.665024, "lr_stage": "cosine_decay", "mfu": 58.110778, "stage_name": "permutation_cycle", "status": "training", "step": 495, "tokens_per_sec_per_gpu": 101161.000218, "vram_gb": 20.709735} {"adam_lr": 9.347474647526095e-05, "dt_ms": 164.35051, "global_tokens_per_sec": 99689.377511, "grad_norm": 41.75, "loss": 3.94118, "lr_stage": "cosine_decay", "mfu": 57.265421, "stage_name": "permutation_cycle", "status": "training", "step": 496, "tokens_per_sec_per_gpu": 99689.377511, "vram_gb": 20.709735} {"adam_lr": 9.343157572190957e-05, "dt_ms": 163.944483, "global_tokens_per_sec": 99936.2694, "grad_norm": 37.25, "loss": 4.405493, "lr_stage": "cosine_decay", "mfu": 57.407245, "stage_name": "permutation_cycle", "status": "training", "step": 497, "tokens_per_sec_per_gpu": 99936.2694, "vram_gb": 20.709735} {"adam_lr": 9.338827266844644e-05, "dt_ms": 163.751125, "global_tokens_per_sec": 100054.274231, "grad_norm": 30.5, "loss": 3.596349, "lr_stage": "cosine_decay", "mfu": 57.475031, "stage_name": "permutation_cycle", "status": "training", "step": 498, "tokens_per_sec_per_gpu": 100054.274231, "vram_gb": 20.709735} {"adam_lr": 9.334483744678014e-05, "dt_ms": 163.376093, "global_tokens_per_sec": 100283.950412, "grad_norm": 39.75, "loss": 2.776109, "lr_stage": "cosine_decay", "mfu": 57.606966, "stage_name": "permutation_cycle", "status": "training", "step": 499, "tokens_per_sec_per_gpu": 100283.950412, "vram_gb": 20.709735} {"adam_lr": 9.330127018922194e-05, "dt_ms": 163.737774, "global_tokens_per_sec": 100062.432817, "grad_norm": 57.25, "loss": 4.799246, "lr_stage": "cosine_decay", "mfu": 57.479718, "stage_name": "permutation_cycle", "status": "training", "step": 500, "tokens_per_sec_per_gpu": 100062.432817, "vram_gb": 20.709735} {"adam_lr": 9.325757102848523e-05, "dt_ms": 162.243128, "global_tokens_per_sec": 100984.24642, "grad_norm": 40.5, "loss": 2.798522, "lr_stage": "cosine_decay", "mfu": 58.009243, "stage_name": "permutation_cycle", "status": "training", "step": 501, "tokens_per_sec_per_gpu": 100984.24642, "vram_gb": 20.709735} {"adam_lr": 9.321374009768525e-05, "dt_ms": 165.119171, "global_tokens_per_sec": 99225.304286, "grad_norm": 28.75, "loss": 3.877379, "lr_stage": "cosine_decay", "mfu": 56.998839, "stage_name": "permutation_cycle", "status": "training", "step": 502, "tokens_per_sec_per_gpu": 99225.304286, "vram_gb": 20.709735} {"adam_lr": 9.316977753033859e-05, "dt_ms": 162.386894, "global_tokens_per_sec": 100894.841779, "grad_norm": 26.375, "loss": 3.221231, "lr_stage": "cosine_decay", "mfu": 57.957886, "stage_name": "permutation_cycle", "status": "training", "step": 503, "tokens_per_sec_per_gpu": 100894.841779, "vram_gb": 20.709735} {"adam_lr": 9.312568346036288e-05, "dt_ms": 162.948608, "global_tokens_per_sec": 100547.038487, "grad_norm": 32.25, "loss": 2.879389, "lr_stage": "cosine_decay", "mfu": 57.758094, "stage_name": "permutation_cycle", "status": "training", "step": 504, "tokens_per_sec_per_gpu": 100547.038487, "vram_gb": 20.709735} {"adam_lr": 9.308145802207629e-05, "dt_ms": 161.767721, "global_tokens_per_sec": 101281.021213, "grad_norm": 32.5, "loss": 2.936154, "lr_stage": "cosine_decay", "mfu": 58.179722, "stage_name": "permutation_cycle", "status": "training", "step": 505, "tokens_per_sec_per_gpu": 101281.021213, "vram_gb": 20.709735} {"adam_lr": 9.30371013501972e-05, "dt_ms": 163.500071, "global_tokens_per_sec": 100207.907817, "grad_norm": 36.0, "loss": 3.572638, "lr_stage": "cosine_decay", "mfu": 57.563284, "stage_name": "permutation_cycle", "status": "training", "step": 506, "tokens_per_sec_per_gpu": 100207.907817, "vram_gb": 20.709735} {"adam_lr": 9.299261357984368e-05, "dt_ms": 161.786556, "global_tokens_per_sec": 101269.230153, "grad_norm": 51.25, "loss": 4.105769, "lr_stage": "cosine_decay", "mfu": 58.172949, "stage_name": "permutation_cycle", "status": "training", "step": 507, "tokens_per_sec_per_gpu": 101269.230153, "vram_gb": 20.709735} {"adam_lr": 9.294799484653323e-05, "dt_ms": 164.948225, "global_tokens_per_sec": 99328.13765, "grad_norm": 45.0, "loss": 4.242117, "lr_stage": "cosine_decay", "mfu": 57.057911, "stage_name": "permutation_cycle", "status": "training", "step": 508, "tokens_per_sec_per_gpu": 99328.13765, "vram_gb": 20.709735} {"adam_lr": 9.290324528618224e-05, "dt_ms": 162.105799, "global_tokens_per_sec": 101069.795956, "grad_norm": 39.0, "loss": 3.168252, "lr_stage": "cosine_decay", "mfu": 58.058386, "stage_name": "permutation_cycle", "status": "training", "step": 509, "tokens_per_sec_per_gpu": 101069.795956, "vram_gb": 20.709735} {"adam_lr": 9.285836503510562e-05, "dt_ms": 164.469481, "global_tokens_per_sec": 99617.266065, "grad_norm": 41.5, "loss": 4.499303, "lr_stage": "cosine_decay", "mfu": 57.223997, "stage_name": "permutation_cycle", "status": "training", "step": 510, "tokens_per_sec_per_gpu": 99617.266065, "vram_gb": 20.709735} {"adam_lr": 9.281335423001642e-05, "dt_ms": 162.5278, "global_tokens_per_sec": 100807.369814, "grad_norm": 38.5, "loss": 3.153351, "lr_stage": "cosine_decay", "mfu": 57.907639, "stage_name": "permutation_cycle", "status": "training", "step": 511, "tokens_per_sec_per_gpu": 100807.369814, "vram_gb": 20.709735} {"adam_lr": 9.276821300802534e-05, "dt_ms": 161.631107, "global_tokens_per_sec": 101366.625958, "grad_norm": 39.75, "loss": 3.017159, "lr_stage": "cosine_decay", "mfu": 58.228897, "stage_name": "permutation_cycle", "status": "training", "step": 512, "tokens_per_sec_per_gpu": 101366.625958, "vram_gb": 20.709735} {"adam_lr": 9.272294150664038e-05, "dt_ms": 161.963701, "global_tokens_per_sec": 101158.46868, "grad_norm": 27.625, "loss": 2.932367, "lr_stage": "cosine_decay", "mfu": 58.109323, "stage_name": "permutation_cycle", "status": "training", "step": 513, "tokens_per_sec_per_gpu": 101158.46868, "vram_gb": 20.709735} {"adam_lr": 9.267753986376637e-05, "dt_ms": 163.439274, "global_tokens_per_sec": 100245.183521, "grad_norm": 43.0, "loss": 3.914881, "lr_stage": "cosine_decay", "mfu": 57.584697, "stage_name": "permutation_cycle", "status": "training", "step": 514, "tokens_per_sec_per_gpu": 100245.183521, "vram_gb": 20.709735} {"adam_lr": 9.263200821770461e-05, "dt_ms": 163.225174, "global_tokens_per_sec": 100376.673545, "grad_norm": 34.25, "loss": 3.221276, "lr_stage": "cosine_decay", "mfu": 57.66023, "stage_name": "permutation_cycle", "status": "training", "step": 515, "tokens_per_sec_per_gpu": 100376.673545, "vram_gb": 20.709735} {"adam_lr": 9.258634670715238e-05, "dt_ms": 163.222551, "global_tokens_per_sec": 100378.286364, "grad_norm": 32.75, "loss": 3.969903, "lr_stage": "cosine_decay", "mfu": 57.661156, "stage_name": "permutation_cycle", "status": "training", "step": 516, "tokens_per_sec_per_gpu": 100378.286364, "vram_gb": 20.709735} {"adam_lr": 9.254055547120257e-05, "dt_ms": 163.35535, "global_tokens_per_sec": 100296.684194, "grad_norm": 37.5, "loss": 3.178524, "lr_stage": "cosine_decay", "mfu": 57.614281, "stage_name": "permutation_cycle", "status": "training", "step": 517, "tokens_per_sec_per_gpu": 100296.684194, "vram_gb": 20.709735} {"adam_lr": 9.249463464934321e-05, "dt_ms": 163.437128, "global_tokens_per_sec": 100246.49964, "grad_norm": 43.75, "loss": 3.515912, "lr_stage": "cosine_decay", "mfu": 57.585453, "stage_name": "permutation_cycle", "status": "training", "step": 518, "tokens_per_sec_per_gpu": 100246.49964, "vram_gb": 20.709735} {"adam_lr": 9.244858438145708e-05, "dt_ms": 165.583611, "global_tokens_per_sec": 98946.990871, "grad_norm": 37.25, "loss": 4.235371, "lr_stage": "cosine_decay", "mfu": 56.838965, "stage_name": "permutation_cycle", "status": "training", "step": 519, "tokens_per_sec_per_gpu": 98946.990871, "vram_gb": 20.709735} {"adam_lr": 9.24024048078213e-05, "dt_ms": 161.974669, "global_tokens_per_sec": 101151.619271, "grad_norm": 54.25, "loss": 5.104409, "lr_stage": "cosine_decay", "mfu": 58.105389, "stage_name": "permutation_cycle", "status": "training", "step": 520, "tokens_per_sec_per_gpu": 101151.619271, "vram_gb": 20.709735} {"adam_lr": 9.235609606910687e-05, "dt_ms": 163.896322, "global_tokens_per_sec": 99965.63544, "grad_norm": 29.125, "loss": 3.946562, "lr_stage": "cosine_decay", "mfu": 57.424114, "stage_name": "permutation_cycle", "status": "training", "step": 521, "tokens_per_sec_per_gpu": 99965.63544, "vram_gb": 20.709735} {"adam_lr": 9.230965830637821e-05, "dt_ms": 163.690329, "global_tokens_per_sec": 100091.435703, "grad_norm": 38.5, "loss": 3.885385, "lr_stage": "cosine_decay", "mfu": 57.496378, "stage_name": "permutation_cycle", "status": "training", "step": 522, "tokens_per_sec_per_gpu": 100091.435703, "vram_gb": 20.709735} {"adam_lr": 9.226309166109281e-05, "dt_ms": 162.552595, "global_tokens_per_sec": 100791.992807, "grad_norm": 41.0, "loss": 3.863533, "lr_stage": "cosine_decay", "mfu": 57.898805, "stage_name": "permutation_cycle", "status": "training", "step": 523, "tokens_per_sec_per_gpu": 100791.992807, "vram_gb": 20.709735} {"adam_lr": 9.221639627510076e-05, "dt_ms": 163.505316, "global_tokens_per_sec": 100204.693173, "grad_norm": 27.125, "loss": 3.231284, "lr_stage": "cosine_decay", "mfu": 57.561438, "stage_name": "permutation_cycle", "status": "training", "step": 524, "tokens_per_sec_per_gpu": 100204.693173, "vram_gb": 20.709735} {"adam_lr": 9.21695722906443e-05, "dt_ms": 164.877415, "global_tokens_per_sec": 99370.796355, "grad_norm": 33.25, "loss": 3.542789, "lr_stage": "cosine_decay", "mfu": 57.082415, "stage_name": "permutation_cycle", "status": "training", "step": 525, "tokens_per_sec_per_gpu": 99370.796355, "vram_gb": 20.709735} {"adam_lr": 9.212261985035739e-05, "dt_ms": 162.726402, "global_tokens_per_sec": 100684.337453, "grad_norm": 34.75, "loss": 3.216183, "lr_stage": "cosine_decay", "mfu": 57.836964, "stage_name": "permutation_cycle", "status": "training", "step": 526, "tokens_per_sec_per_gpu": 100684.337453, "vram_gb": 20.709735} {"adam_lr": 9.207553909726531e-05, "dt_ms": 163.943291, "global_tokens_per_sec": 99936.996073, "grad_norm": 32.5, "loss": 4.003968, "lr_stage": "cosine_decay", "mfu": 57.407662, "stage_name": "permutation_cycle", "status": "training", "step": 527, "tokens_per_sec_per_gpu": 99936.996073, "vram_gb": 20.709735} {"adam_lr": 9.202833017478422e-05, "dt_ms": 163.291454, "global_tokens_per_sec": 100335.93043, "grad_norm": 37.0, "loss": 3.126728, "lr_stage": "cosine_decay", "mfu": 57.636826, "stage_name": "permutation_cycle", "status": "training", "step": 528, "tokens_per_sec_per_gpu": 100335.93043, "vram_gb": 20.709735} {"adam_lr": 9.198099322672067e-05, "dt_ms": 162.807703, "global_tokens_per_sec": 100634.059054, "grad_norm": 35.25, "loss": 3.994625, "lr_stage": "cosine_decay", "mfu": 57.808082, "stage_name": "permutation_cycle", "status": "training", "step": 529, "tokens_per_sec_per_gpu": 100634.059054, "vram_gb": 20.709735} {"adam_lr": 9.193352839727121e-05, "dt_ms": 163.698912, "global_tokens_per_sec": 100086.187704, "grad_norm": 42.75, "loss": 3.510785, "lr_stage": "cosine_decay", "mfu": 57.493364, "stage_name": "permutation_cycle", "status": "training", "step": 530, "tokens_per_sec_per_gpu": 100086.187704, "vram_gb": 20.709735} {"adam_lr": 9.188593583102195e-05, "dt_ms": 165.798903, "global_tokens_per_sec": 98818.506949, "grad_norm": 32.25, "loss": 4.50124, "lr_stage": "cosine_decay", "mfu": 56.765159, "stage_name": "permutation_cycle", "status": "training", "step": 531, "tokens_per_sec_per_gpu": 98818.506949, "vram_gb": 20.709735} {"adam_lr": 9.183821567294809e-05, "dt_ms": 163.134575, "global_tokens_per_sec": 100432.41913, "grad_norm": 20.75, "loss": 2.408216, "lr_stage": "cosine_decay", "mfu": 57.692252, "stage_name": "permutation_cycle", "status": "training", "step": 532, "tokens_per_sec_per_gpu": 100432.41913, "vram_gb": 20.709735} {"adam_lr": 9.179036806841353e-05, "dt_ms": 161.855221, "global_tokens_per_sec": 101226.268263, "grad_norm": 23.625, "loss": 2.080314, "lr_stage": "cosine_decay", "mfu": 58.14827, "stage_name": "permutation_cycle", "status": "training", "step": 533, "tokens_per_sec_per_gpu": 101226.268263, "vram_gb": 20.709735} {"adam_lr": 9.174239316317033e-05, "dt_ms": 166.041613, "global_tokens_per_sec": 98674.059719, "grad_norm": 33.5, "loss": 3.21558, "lr_stage": "cosine_decay", "mfu": 56.682183, "stage_name": "permutation_cycle", "status": "training", "step": 534, "tokens_per_sec_per_gpu": 98674.059719, "vram_gb": 20.709735} {"adam_lr": 9.169429110335841e-05, "dt_ms": 160.949469, "global_tokens_per_sec": 101795.924778, "grad_norm": 51.75, "loss": 4.692546, "lr_stage": "cosine_decay", "mfu": 58.475503, "stage_name": "permutation_cycle", "status": "training", "step": 535, "tokens_per_sec_per_gpu": 101795.924778, "vram_gb": 20.709735} {"adam_lr": 9.164606203550497e-05, "dt_ms": 160.506248, "global_tokens_per_sec": 102077.022893, "grad_norm": 37.0, "loss": 3.58756, "lr_stage": "cosine_decay", "mfu": 58.636976, "stage_name": "permutation_cycle", "status": "training", "step": 536, "tokens_per_sec_per_gpu": 102077.022893, "vram_gb": 20.709735} {"adam_lr": 9.159770610652413e-05, "dt_ms": 164.110422, "global_tokens_per_sec": 99835.219402, "grad_norm": 42.25, "loss": 3.424228, "lr_stage": "cosine_decay", "mfu": 57.349198, "stage_name": "permutation_cycle", "status": "training", "step": 537, "tokens_per_sec_per_gpu": 99835.219402, "vram_gb": 20.709735} {"adam_lr": 9.154922346371642e-05, "dt_ms": 164.678097, "global_tokens_per_sec": 99491.069676, "grad_norm": 42.75, "loss": 3.836627, "lr_stage": "cosine_decay", "mfu": 57.151505, "stage_name": "permutation_cycle", "status": "training", "step": 538, "tokens_per_sec_per_gpu": 99491.069676, "vram_gb": 20.709735} {"adam_lr": 9.150061425476838e-05, "dt_ms": 163.134336, "global_tokens_per_sec": 100432.565911, "grad_norm": 36.25, "loss": 3.342193, "lr_stage": "cosine_decay", "mfu": 57.692337, "stage_name": "permutation_cycle", "status": "training", "step": 539, "tokens_per_sec_per_gpu": 100432.565911, "vram_gb": 20.709735} {"adam_lr": 9.145187862775209e-05, "dt_ms": 165.67111, "global_tokens_per_sec": 98894.731766, "grad_norm": 33.25, "loss": 4.105666, "lr_stage": "cosine_decay", "mfu": 56.808946, "stage_name": "permutation_cycle", "status": "training", "step": 540, "tokens_per_sec_per_gpu": 98894.731766, "vram_gb": 20.709735} {"adam_lr": 9.140301673112472e-05, "dt_ms": 162.06336, "global_tokens_per_sec": 101096.262464, "grad_norm": 46.25, "loss": 3.940114, "lr_stage": "cosine_decay", "mfu": 58.07359, "stage_name": "permutation_cycle", "status": "training", "step": 541, "tokens_per_sec_per_gpu": 101096.262464, "vram_gb": 20.709735} {"adam_lr": 9.135402871372808e-05, "dt_ms": 162.721157, "global_tokens_per_sec": 100687.582946, "grad_norm": 35.5, "loss": 2.870248, "lr_stage": "cosine_decay", "mfu": 57.838828, "stage_name": "permutation_cycle", "status": "training", "step": 542, "tokens_per_sec_per_gpu": 100687.582946, "vram_gb": 20.709735} {"adam_lr": 9.130491472478819e-05, "dt_ms": 162.117958, "global_tokens_per_sec": 101062.215409, "grad_norm": 42.0, "loss": 3.514457, "lr_stage": "cosine_decay", "mfu": 58.054032, "stage_name": "permutation_cycle", "status": "training", "step": 543, "tokens_per_sec_per_gpu": 101062.215409, "vram_gb": 20.709735} {"adam_lr": 9.125567491391476e-05, "dt_ms": 161.372423, "global_tokens_per_sec": 101529.119275, "grad_norm": 42.25, "loss": 2.840641, "lr_stage": "cosine_decay", "mfu": 58.322239, "stage_name": "permutation_cycle", "status": "training", "step": 544, "tokens_per_sec_per_gpu": 101529.119275, "vram_gb": 20.709735} {"adam_lr": 9.120630943110077e-05, "dt_ms": 163.733959, "global_tokens_per_sec": 100064.764086, "grad_norm": 32.75, "loss": 2.995456, "lr_stage": "cosine_decay", "mfu": 57.481057, "stage_name": "permutation_cycle", "status": "training", "step": 545, "tokens_per_sec_per_gpu": 100064.764086, "vram_gb": 20.709735} {"adam_lr": 9.11568184267221e-05, "dt_ms": 164.014101, "global_tokens_per_sec": 99893.849963, "grad_norm": 27.125, "loss": 2.763564, "lr_stage": "cosine_decay", "mfu": 57.382878, "stage_name": "permutation_cycle", "status": "training", "step": 546, "tokens_per_sec_per_gpu": 99893.849963, "vram_gb": 20.709735} {"adam_lr": 9.110720205153687e-05, "dt_ms": 161.428213, "global_tokens_per_sec": 101494.030587, "grad_norm": 30.25, "loss": 3.9389, "lr_stage": "cosine_decay", "mfu": 58.302083, "stage_name": "permutation_cycle", "status": "training", "step": 547, "tokens_per_sec_per_gpu": 101494.030587, "vram_gb": 20.709735} {"adam_lr": 9.105746045668521e-05, "dt_ms": 162.6966, "global_tokens_per_sec": 100702.780537, "grad_norm": 43.25, "loss": 4.976766, "lr_stage": "cosine_decay", "mfu": 57.847558, "stage_name": "permutation_cycle", "status": "training", "step": 548, "tokens_per_sec_per_gpu": 100702.780537, "vram_gb": 20.709735} {"adam_lr": 9.100759379368862e-05, "dt_ms": 162.457228, "global_tokens_per_sec": 100851.160833, "grad_norm": 37.25, "loss": 3.032335, "lr_stage": "cosine_decay", "mfu": 57.932794, "stage_name": "permutation_cycle", "status": "training", "step": 549, "tokens_per_sec_per_gpu": 100851.160833, "vram_gb": 20.709735} {"adam_lr": 9.09576022144496e-05, "dt_ms": 163.715839, "global_tokens_per_sec": 100075.839097, "grad_norm": 28.625, "loss": 3.954906, "lr_stage": "cosine_decay", "mfu": 57.487419, "stage_name": "permutation_cycle", "status": "training", "step": 550, "tokens_per_sec_per_gpu": 100075.839097, "vram_gb": 20.709735} {"adam_lr": 9.090748587125118e-05, "dt_ms": 163.305283, "global_tokens_per_sec": 100327.434238, "grad_norm": 33.75, "loss": 3.540076, "lr_stage": "cosine_decay", "mfu": 57.631945, "stage_name": "permutation_cycle", "status": "training", "step": 551, "tokens_per_sec_per_gpu": 100327.434238, "vram_gb": 20.709735} {"adam_lr": 9.085724491675642e-05, "dt_ms": 163.795471, "global_tokens_per_sec": 100027.185616, "grad_norm": 50.75, "loss": 3.709647, "lr_stage": "cosine_decay", "mfu": 57.459471, "stage_name": "permutation_cycle", "status": "training", "step": 552, "tokens_per_sec_per_gpu": 100027.185616, "vram_gb": 20.709735} {"adam_lr": 9.080687950400802e-05, "dt_ms": 162.328243, "global_tokens_per_sec": 100931.29619, "grad_norm": 33.75, "loss": 3.46342, "lr_stage": "cosine_decay", "mfu": 57.978827, "stage_name": "permutation_cycle", "status": "training", "step": 553, "tokens_per_sec_per_gpu": 100931.29619, "vram_gb": 20.709735} {"adam_lr": 9.075638978642771e-05, "dt_ms": 163.940907, "global_tokens_per_sec": 99938.449453, "grad_norm": 25.25, "loss": 2.875645, "lr_stage": "cosine_decay", "mfu": 57.408497, "stage_name": "permutation_cycle", "status": "training", "step": 554, "tokens_per_sec_per_gpu": 99938.449453, "vram_gb": 20.709735} {"adam_lr": 9.070577591781597e-05, "dt_ms": 163.646698, "global_tokens_per_sec": 100118.121541, "grad_norm": 41.25, "loss": 3.483328, "lr_stage": "cosine_decay", "mfu": 57.511708, "stage_name": "permutation_cycle", "status": "training", "step": 555, "tokens_per_sec_per_gpu": 100118.121541, "vram_gb": 20.709735} {"adam_lr": 9.065503805235138e-05, "dt_ms": 162.521362, "global_tokens_per_sec": 100811.362689, "grad_norm": 33.0, "loss": 3.616224, "lr_stage": "cosine_decay", "mfu": 57.909932, "stage_name": "permutation_cycle", "status": "training", "step": 556, "tokens_per_sec_per_gpu": 100811.362689, "vram_gb": 20.709735} {"adam_lr": 9.060417634459031e-05, "dt_ms": 161.64875, "global_tokens_per_sec": 101355.56241, "grad_norm": 41.0, "loss": 3.169606, "lr_stage": "cosine_decay", "mfu": 58.222542, "stage_name": "permutation_cycle", "status": "training", "step": 557, "tokens_per_sec_per_gpu": 101355.56241, "vram_gb": 20.709735} {"adam_lr": 9.055319094946633e-05, "dt_ms": 162.671328, "global_tokens_per_sec": 100718.425568, "grad_norm": 29.625, "loss": 2.993031, "lr_stage": "cosine_decay", "mfu": 57.856546, "stage_name": "permutation_cycle", "status": "training", "step": 558, "tokens_per_sec_per_gpu": 100718.425568, "vram_gb": 20.709735} {"adam_lr": 9.050208202228981e-05, "dt_ms": 164.75749, "global_tokens_per_sec": 99443.126891, "grad_norm": 30.875, "loss": 2.904016, "lr_stage": "cosine_decay", "mfu": 57.123965, "stage_name": "permutation_cycle", "status": "training", "step": 559, "tokens_per_sec_per_gpu": 99443.126891, "vram_gb": 20.709735} {"adam_lr": 9.045084971874738e-05, "dt_ms": 164.274454, "global_tokens_per_sec": 99735.531541, "grad_norm": 31.75, "loss": 3.077105, "lr_stage": "cosine_decay", "mfu": 57.291933, "stage_name": "permutation_cycle", "status": "training", "step": 560, "tokens_per_sec_per_gpu": 99735.531541, "vram_gb": 20.709735} {"adam_lr": 9.039949419490153e-05, "dt_ms": 161.298752, "global_tokens_per_sec": 101575.491527, "grad_norm": 37.0, "loss": 3.583774, "lr_stage": "cosine_decay", "mfu": 58.348877, "stage_name": "permutation_cycle", "status": "training", "step": 561, "tokens_per_sec_per_gpu": 101575.491527, "vram_gb": 20.709735} {"adam_lr": 9.034801560719011e-05, "dt_ms": 164.530754, "global_tokens_per_sec": 99580.167189, "grad_norm": 27.5, "loss": 3.397772, "lr_stage": "cosine_decay", "mfu": 57.202686, "stage_name": "permutation_cycle", "status": "training", "step": 562, "tokens_per_sec_per_gpu": 99580.167189, "vram_gb": 20.709735} {"adam_lr": 9.029641411242579e-05, "dt_ms": 165.030241, "global_tokens_per_sec": 99278.773996, "grad_norm": 32.5, "loss": 3.438887, "lr_stage": "cosine_decay", "mfu": 57.029554, "stage_name": "permutation_cycle", "status": "training", "step": 563, "tokens_per_sec_per_gpu": 99278.773996, "vram_gb": 20.709735} {"adam_lr": 9.02446898677957e-05, "dt_ms": 159.374237, "global_tokens_per_sec": 102802.060748, "grad_norm": 34.5, "loss": 3.44382, "lr_stage": "cosine_decay", "mfu": 59.053466, "stage_name": "permutation_cycle", "status": "training", "step": 564, "tokens_per_sec_per_gpu": 102802.060748, "vram_gb": 20.709735} {"adam_lr": 9.019284303086087e-05, "dt_ms": 164.756775, "global_tokens_per_sec": 99443.558602, "grad_norm": 38.5, "loss": 3.844892, "lr_stage": "cosine_decay", "mfu": 57.124213, "stage_name": "permutation_cycle", "status": "training", "step": 565, "tokens_per_sec_per_gpu": 99443.558602, "vram_gb": 20.709735} {"adam_lr": 9.014087375955573e-05, "dt_ms": 165.106297, "global_tokens_per_sec": 99233.041643, "grad_norm": 42.5, "loss": 2.729573, "lr_stage": "cosine_decay", "mfu": 57.003284, "stage_name": "permutation_cycle", "status": "training", "step": 566, "tokens_per_sec_per_gpu": 99233.041643, "vram_gb": 20.709735} {"adam_lr": 9.00887822121877e-05, "dt_ms": 162.248611, "global_tokens_per_sec": 100980.833386, "grad_norm": 36.5, "loss": 3.776759, "lr_stage": "cosine_decay", "mfu": 58.007283, "stage_name": "permutation_cycle", "status": "training", "step": 567, "tokens_per_sec_per_gpu": 100980.833386, "vram_gb": 20.709735} {"adam_lr": 9.003656854743667e-05, "dt_ms": 163.945675, "global_tokens_per_sec": 99935.542736, "grad_norm": 35.75, "loss": 3.317199, "lr_stage": "cosine_decay", "mfu": 57.406827, "stage_name": "permutation_cycle", "status": "training", "step": 568, "tokens_per_sec_per_gpu": 99935.542736, "vram_gb": 20.709735} {"adam_lr": 8.998423292435454e-05, "dt_ms": 163.184404, "global_tokens_per_sec": 100401.751399, "grad_norm": 25.5, "loss": 2.913069, "lr_stage": "cosine_decay", "mfu": 57.674636, "stage_name": "permutation_cycle", "status": "training", "step": 569, "tokens_per_sec_per_gpu": 100401.751399, "vram_gb": 20.709735} {"adam_lr": 8.993177550236464e-05, "dt_ms": 163.348913, "global_tokens_per_sec": 100300.636715, "grad_norm": 43.25, "loss": 3.703437, "lr_stage": "cosine_decay", "mfu": 57.616552, "stage_name": "permutation_cycle", "status": "training", "step": 570, "tokens_per_sec_per_gpu": 100300.636715, "vram_gb": 20.709735} {"adam_lr": 8.987919644126144e-05, "dt_ms": 162.068605, "global_tokens_per_sec": 101092.990572, "grad_norm": 41.0, "loss": 3.856283, "lr_stage": "cosine_decay", "mfu": 58.07171, "stage_name": "permutation_cycle", "status": "training", "step": 571, "tokens_per_sec_per_gpu": 101092.990572, "vram_gb": 20.709735} {"adam_lr": 8.982649590120982e-05, "dt_ms": 164.429188, "global_tokens_per_sec": 99641.676893, "grad_norm": 41.0, "loss": 1.970686, "lr_stage": "cosine_decay", "mfu": 57.23802, "stage_name": "permutation_cycle", "status": "training", "step": 572, "tokens_per_sec_per_gpu": 99641.676893, "vram_gb": 20.709735} {"adam_lr": 8.977367404274479e-05, "dt_ms": 163.706779, "global_tokens_per_sec": 100081.377522, "grad_norm": 41.0, "loss": 3.619716, "lr_stage": "cosine_decay", "mfu": 57.490601, "stage_name": "permutation_cycle", "status": "training", "step": 573, "tokens_per_sec_per_gpu": 100081.377522, "vram_gb": 20.709735} {"adam_lr": 8.972073102677091e-05, "dt_ms": 163.032532, "global_tokens_per_sec": 100495.280453, "grad_norm": 43.75, "loss": 4.454069, "lr_stage": "cosine_decay", "mfu": 57.728362, "stage_name": "permutation_cycle", "status": "training", "step": 574, "tokens_per_sec_per_gpu": 100495.280453, "vram_gb": 20.709735} {"adam_lr": 8.966766701456177e-05, "dt_ms": 161.46636, "global_tokens_per_sec": 101470.05228, "grad_norm": 26.75, "loss": 2.94259, "lr_stage": "cosine_decay", "mfu": 58.288309, "stage_name": "permutation_cycle", "status": "training", "step": 575, "tokens_per_sec_per_gpu": 101470.05228, "vram_gb": 20.709735} {"adam_lr": 8.961448216775954e-05, "dt_ms": 164.132595, "global_tokens_per_sec": 99821.732507, "grad_norm": 54.75, "loss": 4.127586, "lr_stage": "cosine_decay", "mfu": 57.341451, "stage_name": "permutation_cycle", "status": "training", "step": 576, "tokens_per_sec_per_gpu": 99821.732507, "vram_gb": 20.709735} {"adam_lr": 8.956117664837451e-05, "dt_ms": 163.893223, "global_tokens_per_sec": 99967.525925, "grad_norm": 37.0, "loss": 4.408029, "lr_stage": "cosine_decay", "mfu": 57.4252, "stage_name": "permutation_cycle", "status": "training", "step": 577, "tokens_per_sec_per_gpu": 99967.525925, "vram_gb": 20.709735} {"adam_lr": 8.950775061878453e-05, "dt_ms": 163.390636, "global_tokens_per_sec": 100275.024056, "grad_norm": 48.25, "loss": 3.917449, "lr_stage": "cosine_decay", "mfu": 57.601839, "stage_name": "permutation_cycle", "status": "training", "step": 578, "tokens_per_sec_per_gpu": 100275.024056, "vram_gb": 20.709735} {"adam_lr": 8.945420424173454e-05, "dt_ms": 164.993286, "global_tokens_per_sec": 99301.010265, "grad_norm": 37.25, "loss": 3.282534, "lr_stage": "cosine_decay", "mfu": 57.042328, "stage_name": "permutation_cycle", "status": "training", "step": 579, "tokens_per_sec_per_gpu": 99301.010265, "vram_gb": 20.709735} {"adam_lr": 8.940053768033609e-05, "dt_ms": 162.889242, "global_tokens_per_sec": 100583.683622, "grad_norm": 56.5, "loss": 2.585916, "lr_stage": "cosine_decay", "mfu": 57.779145, "stage_name": "permutation_cycle", "status": "training", "step": 580, "tokens_per_sec_per_gpu": 100583.683622, "vram_gb": 20.709735} {"adam_lr": 8.934675109806688e-05, "dt_ms": 161.144018, "global_tokens_per_sec": 101673.026314, "grad_norm": 30.375, "loss": 2.389954, "lr_stage": "cosine_decay", "mfu": 58.404905, "stage_name": "permutation_cycle", "status": "training", "step": 581, "tokens_per_sec_per_gpu": 101673.026314, "vram_gb": 20.709735} {"adam_lr": 8.92928446587701e-05, "dt_ms": 161.412001, "global_tokens_per_sec": 101504.2248, "grad_norm": 36.75, "loss": 3.391608, "lr_stage": "cosine_decay", "mfu": 58.307939, "stage_name": "permutation_cycle", "status": "training", "step": 582, "tokens_per_sec_per_gpu": 101504.2248, "vram_gb": 20.709735} {"adam_lr": 8.923881852665415e-05, "dt_ms": 164.315224, "global_tokens_per_sec": 99710.785353, "grad_norm": 31.5, "loss": 3.252412, "lr_stage": "cosine_decay", "mfu": 57.277718, "stage_name": "permutation_cycle", "status": "training", "step": 583, "tokens_per_sec_per_gpu": 99710.785353, "vram_gb": 20.709735} {"adam_lr": 8.9184672866292e-05, "dt_ms": 164.628983, "global_tokens_per_sec": 99520.751127, "grad_norm": 37.25, "loss": 3.442295, "lr_stage": "cosine_decay", "mfu": 57.168555, "stage_name": "permutation_cycle", "status": "training", "step": 584, "tokens_per_sec_per_gpu": 99520.751127, "vram_gb": 20.709735} {"adam_lr": 8.91304078426207e-05, "dt_ms": 164.415121, "global_tokens_per_sec": 99650.201834, "grad_norm": 30.875, "loss": 4.422324, "lr_stage": "cosine_decay", "mfu": 57.242917, "stage_name": "permutation_cycle", "status": "training", "step": 585, "tokens_per_sec_per_gpu": 99650.201834, "vram_gb": 20.709735} {"adam_lr": 8.907602362094094e-05, "dt_ms": 164.227009, "global_tokens_per_sec": 99764.345206, "grad_norm": 34.5, "loss": 2.759158, "lr_stage": "cosine_decay", "mfu": 57.308485, "stage_name": "permutation_cycle", "status": "training", "step": 586, "tokens_per_sec_per_gpu": 99764.345206, "vram_gb": 20.709735} {"adam_lr": 8.90215203669165e-05, "dt_ms": 162.035704, "global_tokens_per_sec": 101113.517762, "grad_norm": 22.625, "loss": 2.127026, "lr_stage": "cosine_decay", "mfu": 58.083502, "stage_name": "permutation_cycle", "status": "training", "step": 587, "tokens_per_sec_per_gpu": 101113.517762, "vram_gb": 20.709735} {"adam_lr": 8.896689824657372e-05, "dt_ms": 159.852982, "global_tokens_per_sec": 102494.178334, "grad_norm": 35.75, "loss": 3.285395, "lr_stage": "cosine_decay", "mfu": 58.876606, "stage_name": "permutation_cycle", "status": "training", "step": 588, "tokens_per_sec_per_gpu": 102494.178334, "vram_gb": 20.709735} {"adam_lr": 8.891215742630106e-05, "dt_ms": 161.6714, "global_tokens_per_sec": 101341.362745, "grad_norm": 34.25, "loss": 3.893293, "lr_stage": "cosine_decay", "mfu": 58.214385, "stage_name": "permutation_cycle", "status": "training", "step": 589, "tokens_per_sec_per_gpu": 101341.362745, "vram_gb": 20.709735} {"adam_lr": 8.885729807284856e-05, "dt_ms": 162.463188, "global_tokens_per_sec": 100847.460797, "grad_norm": 40.25, "loss": 2.891946, "lr_stage": "cosine_decay", "mfu": 57.930668, "stage_name": "permutation_cycle", "status": "training", "step": 590, "tokens_per_sec_per_gpu": 100847.460797, "vram_gb": 20.709735} {"adam_lr": 8.88023203533273e-05, "dt_ms": 163.112164, "global_tokens_per_sec": 100446.218382, "grad_norm": 38.5, "loss": 3.442872, "lr_stage": "cosine_decay", "mfu": 57.700179, "stage_name": "permutation_cycle", "status": "training", "step": 591, "tokens_per_sec_per_gpu": 100446.218382, "vram_gb": 20.709735} {"adam_lr": 8.874722443520899e-05, "dt_ms": 161.19194, "global_tokens_per_sec": 101642.799068, "grad_norm": 45.25, "loss": 2.598329, "lr_stage": "cosine_decay", "mfu": 58.387541, "stage_name": "permutation_cycle", "status": "training", "step": 592, "tokens_per_sec_per_gpu": 101642.799068, "vram_gb": 20.709735} {"adam_lr": 8.869201048632532e-05, "dt_ms": 160.979986, "global_tokens_per_sec": 101776.626944, "grad_norm": 32.5, "loss": 2.568498, "lr_stage": "cosine_decay", "mfu": 58.464417, "stage_name": "permutation_cycle", "status": "training", "step": 593, "tokens_per_sec_per_gpu": 101776.626944, "vram_gb": 20.709735} {"adam_lr": 8.863667867486756e-05, "dt_ms": 164.349318, "global_tokens_per_sec": 99690.100599, "grad_norm": 36.5, "loss": 2.844894, "lr_stage": "cosine_decay", "mfu": 57.265836, "stage_name": "permutation_cycle", "status": "training", "step": 594, "tokens_per_sec_per_gpu": 99690.100599, "vram_gb": 20.709735} {"adam_lr": 8.858122916938601e-05, "dt_ms": 160.250902, "global_tokens_per_sec": 102239.674021, "grad_norm": 34.5, "loss": 2.10964, "lr_stage": "cosine_decay", "mfu": 58.730409, "stage_name": "permutation_cycle", "status": "training", "step": 595, "tokens_per_sec_per_gpu": 102239.674021, "vram_gb": 20.709735} {"adam_lr": 8.852566213878947e-05, "dt_ms": 163.746357, "global_tokens_per_sec": 100057.187859, "grad_norm": 37.5, "loss": 4.477349, "lr_stage": "cosine_decay", "mfu": 57.476705, "stage_name": "permutation_cycle", "status": "training", "step": 596, "tokens_per_sec_per_gpu": 100057.187859, "vram_gb": 20.709735} {"adam_lr": 8.846997775234476e-05, "dt_ms": 164.045334, "global_tokens_per_sec": 99874.831025, "grad_norm": 32.25, "loss": 3.945392, "lr_stage": "cosine_decay", "mfu": 57.371952, "stage_name": "permutation_cycle", "status": "training", "step": 597, "tokens_per_sec_per_gpu": 99874.831025, "vram_gb": 20.709735} {"adam_lr": 8.841417617967618e-05, "dt_ms": 163.839579, "global_tokens_per_sec": 100000.257185, "grad_norm": 31.25, "loss": 3.174912, "lr_stage": "cosine_decay", "mfu": 57.444002, "stage_name": "permutation_cycle", "status": "training", "step": 598, "tokens_per_sec_per_gpu": 100000.257185, "vram_gb": 20.709735} {"adam_lr": 8.8358257590765e-05, "dt_ms": 162.886858, "global_tokens_per_sec": 100585.155872, "grad_norm": 46.25, "loss": 4.285923, "lr_stage": "cosine_decay", "mfu": 57.77999, "stage_name": "permutation_cycle", "status": "training", "step": 599, "tokens_per_sec_per_gpu": 100585.155872, "vram_gb": 20.709735} {"adam_lr": 8.83022221559489e-05, "dt_ms": 163.063049, "global_tokens_per_sec": 100476.472559, "grad_norm": 30.625, "loss": 3.396294, "lr_stage": "cosine_decay", "mfu": 57.717558, "stage_name": "permutation_cycle", "status": "training", "step": 600, "tokens_per_sec_per_gpu": 100476.472559, "vram_gb": 20.709735} {"adam_lr": 8.82460700459216e-05, "dt_ms": 162.497282, "global_tokens_per_sec": 100826.301803, "grad_norm": 28.75, "loss": 3.356485, "lr_stage": "cosine_decay", "mfu": 57.918514, "stage_name": "permutation_cycle", "status": "training", "step": 601, "tokens_per_sec_per_gpu": 100826.301803, "vram_gb": 20.709735} {"adam_lr": 8.818980143173213e-05, "dt_ms": 164.963007, "global_tokens_per_sec": 99319.237086, "grad_norm": 37.75, "loss": 3.56096, "lr_stage": "cosine_decay", "mfu": 57.052798, "stage_name": "permutation_cycle", "status": "training", "step": 602, "tokens_per_sec_per_gpu": 99319.237086, "vram_gb": 20.709735} {"adam_lr": 8.813341648478441e-05, "dt_ms": 163.546562, "global_tokens_per_sec": 100179.421567, "grad_norm": 37.75, "loss": 3.927952, "lr_stage": "cosine_decay", "mfu": 57.546921, "stage_name": "permutation_cycle", "status": "training", "step": 603, "tokens_per_sec_per_gpu": 100179.421567, "vram_gb": 20.709735} {"adam_lr": 8.807691537683685e-05, "dt_ms": 162.708282, "global_tokens_per_sec": 100695.550043, "grad_norm": 46.25, "loss": 2.966766, "lr_stage": "cosine_decay", "mfu": 57.843405, "stage_name": "permutation_cycle", "status": "training", "step": 604, "tokens_per_sec_per_gpu": 100695.550043, "vram_gb": 20.709735} {"adam_lr": 8.802029828000156e-05, "dt_ms": 163.257599, "global_tokens_per_sec": 100356.737528, "grad_norm": 41.25, "loss": 3.553572, "lr_stage": "cosine_decay", "mfu": 57.648778, "stage_name": "permutation_cycle", "status": "training", "step": 605, "tokens_per_sec_per_gpu": 100356.737528, "vram_gb": 20.709735} {"adam_lr": 8.796356536674403e-05, "dt_ms": 163.082361, "global_tokens_per_sec": 100464.574325, "grad_norm": 43.0, "loss": 3.164186, "lr_stage": "cosine_decay", "mfu": 57.710724, "stage_name": "permutation_cycle", "status": "training", "step": 606, "tokens_per_sec_per_gpu": 100464.574325, "vram_gb": 20.709735} {"adam_lr": 8.790671680988262e-05, "dt_ms": 163.895845, "global_tokens_per_sec": 99965.926279, "grad_norm": 35.5, "loss": 3.06333, "lr_stage": "cosine_decay", "mfu": 57.424281, "stage_name": "permutation_cycle", "status": "training", "step": 607, "tokens_per_sec_per_gpu": 99965.926279, "vram_gb": 20.709735} {"adam_lr": 8.784975278258783e-05, "dt_ms": 165.18116, "global_tokens_per_sec": 99188.067227, "grad_norm": 32.75, "loss": 2.647839, "lr_stage": "cosine_decay", "mfu": 56.977449, "stage_name": "permutation_cycle", "status": "training", "step": 608, "tokens_per_sec_per_gpu": 99188.067227, "vram_gb": 20.709735} {"adam_lr": 8.779267345838199e-05, "dt_ms": 165.195465, "global_tokens_per_sec": 99179.478028, "grad_norm": 35.25, "loss": 3.99075, "lr_stage": "cosine_decay", "mfu": 56.972515, "stage_name": "permutation_cycle", "status": "training", "step": 609, "tokens_per_sec_per_gpu": 99179.478028, "vram_gb": 20.709735} {"adam_lr": 8.773547901113862e-05, "dt_ms": 163.408756, "global_tokens_per_sec": 100263.904918, "grad_norm": 33.75, "loss": 3.406127, "lr_stage": "cosine_decay", "mfu": 57.595451, "stage_name": "permutation_cycle", "status": "training", "step": 610, "tokens_per_sec_per_gpu": 100263.904918, "vram_gb": 20.709735} {"adam_lr": 8.76781696150819e-05, "dt_ms": 162.541389, "global_tokens_per_sec": 100798.941451, "grad_norm": 29.0, "loss": 3.295637, "lr_stage": "cosine_decay", "mfu": 57.902797, "stage_name": "permutation_cycle", "status": "training", "step": 611, "tokens_per_sec_per_gpu": 100798.941451, "vram_gb": 20.709735} {"adam_lr": 8.762074544478623e-05, "dt_ms": 162.698507, "global_tokens_per_sec": 100701.599978, "grad_norm": 23.75, "loss": 2.98882, "lr_stage": "cosine_decay", "mfu": 57.84688, "stage_name": "permutation_cycle", "status": "training", "step": 612, "tokens_per_sec_per_gpu": 100701.599978, "vram_gb": 20.709735} {"adam_lr": 8.756320667517556e-05, "dt_ms": 162.998438, "global_tokens_per_sec": 100516.300726, "grad_norm": 27.625, "loss": 4.13253, "lr_stage": "cosine_decay", "mfu": 57.740437, "stage_name": "permutation_cycle", "status": "training", "step": 613, "tokens_per_sec_per_gpu": 100516.300726, "vram_gb": 20.709735} {"adam_lr": 8.750555348152298e-05, "dt_ms": 163.761616, "global_tokens_per_sec": 100047.864847, "grad_norm": 34.5, "loss": 2.996953, "lr_stage": "cosine_decay", "mfu": 57.47135, "stage_name": "permutation_cycle", "status": "training", "step": 614, "tokens_per_sec_per_gpu": 100047.864847, "vram_gb": 20.709735} {"adam_lr": 8.744778603945011e-05, "dt_ms": 164.293528, "global_tokens_per_sec": 99723.952848, "grad_norm": 60.0, "loss": 4.494295, "lr_stage": "cosine_decay", "mfu": 57.285282, "stage_name": "permutation_cycle", "status": "training", "step": 615, "tokens_per_sec_per_gpu": 99723.952848, "vram_gb": 20.709735} {"adam_lr": 8.73899045249266e-05, "dt_ms": 162.724495, "global_tokens_per_sec": 100685.517608, "grad_norm": 56.0, "loss": 3.764053, "lr_stage": "cosine_decay", "mfu": 57.837642, "stage_name": "permutation_cycle", "status": "training", "step": 616, "tokens_per_sec_per_gpu": 100685.517608, "vram_gb": 20.709735} {"adam_lr": 8.733190911426958e-05, "dt_ms": 164.579868, "global_tokens_per_sec": 99550.450292, "grad_norm": 26.25, "loss": 3.419818, "lr_stage": "cosine_decay", "mfu": 57.185616, "stage_name": "permutation_cycle", "status": "training", "step": 617, "tokens_per_sec_per_gpu": 99550.450292, "vram_gb": 20.709735} {"adam_lr": 8.727379998414311e-05, "dt_ms": 162.885427, "global_tokens_per_sec": 100586.039242, "grad_norm": 33.0, "loss": 2.96482, "lr_stage": "cosine_decay", "mfu": 57.780498, "stage_name": "permutation_cycle", "status": "training", "step": 618, "tokens_per_sec_per_gpu": 100586.039242, "vram_gb": 20.709735} {"adam_lr": 8.72155773115577e-05, "dt_ms": 160.539627, "global_tokens_per_sec": 102055.799546, "grad_norm": 22.5, "loss": 2.058546, "lr_stage": "cosine_decay", "mfu": 58.624785, "stage_name": "permutation_cycle", "status": "training", "step": 619, "tokens_per_sec_per_gpu": 102055.799546, "vram_gb": 20.709735} {"adam_lr": 8.715724127386972e-05, "dt_ms": 162.586212, "global_tokens_per_sec": 100771.152624, "grad_norm": 33.0, "loss": 2.834247, "lr_stage": "cosine_decay", "mfu": 57.886834, "stage_name": "permutation_cycle", "status": "training", "step": 620, "tokens_per_sec_per_gpu": 100771.152624, "vram_gb": 20.709735} {"adam_lr": 8.709879204878082e-05, "dt_ms": 166.236877, "global_tokens_per_sec": 98558.1554, "grad_norm": 43.75, "loss": 3.678149, "lr_stage": "cosine_decay", "mfu": 56.615603, "stage_name": "permutation_cycle", "status": "training", "step": 621, "tokens_per_sec_per_gpu": 98558.1554, "vram_gb": 20.709735} {"adam_lr": 8.70402298143375e-05, "dt_ms": 164.041519, "global_tokens_per_sec": 99877.153561, "grad_norm": 37.25, "loss": 3.017939, "lr_stage": "cosine_decay", "mfu": 57.373286, "stage_name": "permutation_cycle", "status": "training", "step": 622, "tokens_per_sec_per_gpu": 99877.153561, "vram_gb": 20.709735} {"adam_lr": 8.69815547489305e-05, "dt_ms": 163.496017, "global_tokens_per_sec": 100210.392002, "grad_norm": 37.0, "loss": 3.096789, "lr_stage": "cosine_decay", "mfu": 57.564712, "stage_name": "permutation_cycle", "status": "training", "step": 623, "tokens_per_sec_per_gpu": 100210.392002, "vram_gb": 20.709735} {"adam_lr": 8.692276703129421e-05, "dt_ms": 163.371801, "global_tokens_per_sec": 100286.584723, "grad_norm": 49.75, "loss": 2.538187, "lr_stage": "cosine_decay", "mfu": 57.60848, "stage_name": "permutation_cycle", "status": "training", "step": 624, "tokens_per_sec_per_gpu": 100286.584723, "vram_gb": 20.709735} {"adam_lr": 8.68638668405062e-05, "dt_ms": 163.499355, "global_tokens_per_sec": 100208.346194, "grad_norm": 35.0, "loss": 3.747123, "lr_stage": "cosine_decay", "mfu": 57.563536, "stage_name": "permutation_cycle", "status": "training", "step": 625, "tokens_per_sec_per_gpu": 100208.346194, "vram_gb": 20.709735} {"adam_lr": 8.680485435598673e-05, "dt_ms": 162.493706, "global_tokens_per_sec": 100828.520861, "grad_norm": 50.75, "loss": 3.290126, "lr_stage": "cosine_decay", "mfu": 57.919789, "stage_name": "permutation_cycle", "status": "training", "step": 626, "tokens_per_sec_per_gpu": 100828.520861, "vram_gb": 20.709735} {"adam_lr": 8.6745729757498e-05, "dt_ms": 161.767721, "global_tokens_per_sec": 101281.021213, "grad_norm": 24.125, "loss": 2.782594, "lr_stage": "cosine_decay", "mfu": 58.179722, "stage_name": "permutation_cycle", "status": "training", "step": 627, "tokens_per_sec_per_gpu": 101281.021213, "vram_gb": 20.709735} {"adam_lr": 8.668649322514382e-05, "dt_ms": 162.699461, "global_tokens_per_sec": 100701.009708, "grad_norm": 24.0, "loss": 3.331192, "lr_stage": "cosine_decay", "mfu": 57.846541, "stage_name": "permutation_cycle", "status": "training", "step": 628, "tokens_per_sec_per_gpu": 100701.009708, "vram_gb": 20.709735} {"adam_lr": 8.662714493936895e-05, "dt_ms": 164.209604, "global_tokens_per_sec": 99774.919217, "grad_norm": 25.25, "loss": 2.518935, "lr_stage": "cosine_decay", "mfu": 57.314559, "stage_name": "permutation_cycle", "status": "training", "step": 629, "tokens_per_sec_per_gpu": 99774.919217, "vram_gb": 20.709735} {"adam_lr": 8.656768508095853e-05, "dt_ms": 164.139986, "global_tokens_per_sec": 99817.237685, "grad_norm": 23.75, "loss": 4.197875, "lr_stage": "cosine_decay", "mfu": 57.338869, "stage_name": "permutation_cycle", "status": "training", "step": 630, "tokens_per_sec_per_gpu": 99817.237685, "vram_gb": 20.709735} {"adam_lr": 8.650811383103763e-05, "dt_ms": 165.019512, "global_tokens_per_sec": 99285.228661, "grad_norm": 41.5, "loss": 4.827647, "lr_stage": "cosine_decay", "mfu": 57.033262, "stage_name": "permutation_cycle", "status": "training", "step": 631, "tokens_per_sec_per_gpu": 99285.228661, "vram_gb": 20.709735} {"adam_lr": 8.644843137107059e-05, "dt_ms": 164.809942, "global_tokens_per_sec": 99411.478317, "grad_norm": 28.0, "loss": 2.640562, "lr_stage": "cosine_decay", "mfu": 57.105785, "stage_name": "permutation_cycle", "status": "training", "step": 632, "tokens_per_sec_per_gpu": 99411.478317, "vram_gb": 20.709735} {"adam_lr": 8.638863788286053e-05, "dt_ms": 160.168886, "global_tokens_per_sec": 102292.026812, "grad_norm": 44.75, "loss": 3.43722, "lr_stage": "cosine_decay", "mfu": 58.760483, "stage_name": "permutation_cycle", "status": "training", "step": 633, "tokens_per_sec_per_gpu": 102292.026812, "vram_gb": 20.709735} {"adam_lr": 8.63287335485488e-05, "dt_ms": 163.975477, "global_tokens_per_sec": 99917.379586, "grad_norm": 28.0, "loss": 3.405022, "lr_stage": "cosine_decay", "mfu": 57.396394, "stage_name": "permutation_cycle", "status": "training", "step": 634, "tokens_per_sec_per_gpu": 99917.379586, "vram_gb": 20.709735} {"adam_lr": 8.626871855061438e-05, "dt_ms": 164.677143, "global_tokens_per_sec": 99491.645846, "grad_norm": 35.5, "loss": 3.228039, "lr_stage": "cosine_decay", "mfu": 57.151836, "stage_name": "permutation_cycle", "status": "training", "step": 635, "tokens_per_sec_per_gpu": 99491.645846, "vram_gb": 20.709735} {"adam_lr": 8.620859307187339e-05, "dt_ms": 163.40971, "global_tokens_per_sec": 100263.319768, "grad_norm": 36.25, "loss": 2.508913, "lr_stage": "cosine_decay", "mfu": 57.595115, "stage_name": "permutation_cycle", "status": "training", "step": 636, "tokens_per_sec_per_gpu": 100263.319768, "vram_gb": 20.709735} {"adam_lr": 8.614835729547842e-05, "dt_ms": 163.689613, "global_tokens_per_sec": 100091.873061, "grad_norm": 33.0, "loss": 3.616641, "lr_stage": "cosine_decay", "mfu": 57.49663, "stage_name": "permutation_cycle", "status": "training", "step": 637, "tokens_per_sec_per_gpu": 100091.873061, "vram_gb": 20.709735} {"adam_lr": 8.608801140491811e-05, "dt_ms": 163.733482, "global_tokens_per_sec": 100065.055502, "grad_norm": 34.25, "loss": 3.366455, "lr_stage": "cosine_decay", "mfu": 57.481225, "stage_name": "permutation_cycle", "status": "training", "step": 638, "tokens_per_sec_per_gpu": 100065.055502, "vram_gb": 20.709735} {"adam_lr": 8.602755558401653e-05, "dt_ms": 160.573483, "global_tokens_per_sec": 102034.282022, "grad_norm": 36.0, "loss": 2.623061, "lr_stage": "cosine_decay", "mfu": 58.612424, "stage_name": "permutation_cycle", "status": "training", "step": 639, "tokens_per_sec_per_gpu": 102034.282022, "vram_gb": 20.709735} {"adam_lr": 8.596699001693255e-05, "dt_ms": 161.95631, "global_tokens_per_sec": 101163.085109, "grad_norm": 40.5, "loss": 3.998292, "lr_stage": "cosine_decay", "mfu": 58.111975, "stage_name": "permutation_cycle", "status": "training", "step": 640, "tokens_per_sec_per_gpu": 101163.085109, "vram_gb": 20.709735} {"adam_lr": 8.590631488815944e-05, "dt_ms": 162.689447, "global_tokens_per_sec": 100707.207883, "grad_norm": 31.0, "loss": 3.239415, "lr_stage": "cosine_decay", "mfu": 57.850102, "stage_name": "permutation_cycle", "status": "training", "step": 641, "tokens_per_sec_per_gpu": 100707.207883, "vram_gb": 20.709735} {"adam_lr": 8.584553038252414e-05, "dt_ms": 164.914608, "global_tokens_per_sec": 99348.385195, "grad_norm": 25.625, "loss": 3.25982, "lr_stage": "cosine_decay", "mfu": 57.069542, "stage_name": "permutation_cycle", "status": "training", "step": 642, "tokens_per_sec_per_gpu": 99348.385195, "vram_gb": 20.709735} {"adam_lr": 8.57846366851868e-05, "dt_ms": 164.144516, "global_tokens_per_sec": 99814.482994, "grad_norm": 27.625, "loss": 2.798972, "lr_stage": "cosine_decay", "mfu": 57.337286, "stage_name": "permutation_cycle", "status": "training", "step": 643, "tokens_per_sec_per_gpu": 99814.482994, "vram_gb": 20.709735} {"adam_lr": 8.572363398164017e-05, "dt_ms": 164.480209, "global_tokens_per_sec": 99610.768157, "grad_norm": 39.75, "loss": 3.592865, "lr_stage": "cosine_decay", "mfu": 57.220264, "stage_name": "permutation_cycle", "status": "training", "step": 644, "tokens_per_sec_per_gpu": 99610.768157, "vram_gb": 20.709735} {"adam_lr": 8.566252245770909e-05, "dt_ms": 161.242485, "global_tokens_per_sec": 101610.937063, "grad_norm": 45.25, "loss": 3.913404, "lr_stage": "cosine_decay", "mfu": 58.369239, "stage_name": "permutation_cycle", "status": "training", "step": 645, "tokens_per_sec_per_gpu": 101610.937063, "vram_gb": 20.709735} {"adam_lr": 8.560130229954984e-05, "dt_ms": 162.308693, "global_tokens_per_sec": 100943.453515, "grad_norm": 41.75, "loss": 3.724246, "lr_stage": "cosine_decay", "mfu": 57.98581, "stage_name": "permutation_cycle", "status": "training", "step": 646, "tokens_per_sec_per_gpu": 100943.453515, "vram_gb": 20.709735} {"adam_lr": 8.553997369364963e-05, "dt_ms": 162.487507, "global_tokens_per_sec": 100832.36746, "grad_norm": 40.5, "loss": 4.663138, "lr_stage": "cosine_decay", "mfu": 57.921998, "stage_name": "permutation_cycle", "status": "training", "step": 647, "tokens_per_sec_per_gpu": 100832.36746, "vram_gb": 20.709735} {"adam_lr": 8.547853682682604e-05, "dt_ms": 165.254116, "global_tokens_per_sec": 99144.27786, "grad_norm": 50.5, "loss": 3.919456, "lr_stage": "cosine_decay", "mfu": 56.952294, "stage_name": "permutation_cycle", "status": "training", "step": 648, "tokens_per_sec_per_gpu": 99144.27786, "vram_gb": 20.709735} {"adam_lr": 8.541699188622644e-05, "dt_ms": 163.929224, "global_tokens_per_sec": 99945.571624, "grad_norm": 48.5, "loss": 3.490438, "lr_stage": "cosine_decay", "mfu": 57.412588, "stage_name": "permutation_cycle", "status": "training", "step": 649, "tokens_per_sec_per_gpu": 99945.571624, "vram_gb": 20.709735} {"adam_lr": 8.535533905932738e-05, "dt_ms": 163.289309, "global_tokens_per_sec": 100337.248934, "grad_norm": 33.25, "loss": 3.056774, "lr_stage": "cosine_decay", "mfu": 57.637583, "stage_name": "permutation_cycle", "status": "training", "step": 650, "tokens_per_sec_per_gpu": 100337.248934, "vram_gb": 20.709735} {"adam_lr": 8.529357853393405e-05, "dt_ms": 163.744688, "global_tokens_per_sec": 100058.207669, "grad_norm": 27.75, "loss": 3.409817, "lr_stage": "cosine_decay", "mfu": 57.477291, "stage_name": "permutation_cycle", "status": "training", "step": 651, "tokens_per_sec_per_gpu": 100058.207669, "vram_gb": 20.709735} {"adam_lr": 8.523171049817974e-05, "dt_ms": 163.755655, "global_tokens_per_sec": 100051.506442, "grad_norm": 31.125, "loss": 3.442268, "lr_stage": "cosine_decay", "mfu": 57.473442, "stage_name": "permutation_cycle", "status": "training", "step": 652, "tokens_per_sec_per_gpu": 100051.506442, "vram_gb": 20.709735} {"adam_lr": 8.516973514052519e-05, "dt_ms": 162.535429, "global_tokens_per_sec": 100802.637928, "grad_norm": 30.5, "loss": 3.018513, "lr_stage": "cosine_decay", "mfu": 57.90492, "stage_name": "permutation_cycle", "status": "training", "step": 653, "tokens_per_sec_per_gpu": 100802.637928, "vram_gb": 20.709735} {"adam_lr": 8.510765264975813e-05, "dt_ms": 162.969589, "global_tokens_per_sec": 100534.093981, "grad_norm": 32.25, "loss": 3.542937, "lr_stage": "cosine_decay", "mfu": 57.750658, "stage_name": "permutation_cycle", "status": "training", "step": 654, "tokens_per_sec_per_gpu": 100534.093981, "vram_gb": 20.709735} {"adam_lr": 8.504546321499255e-05, "dt_ms": 163.161755, "global_tokens_per_sec": 100415.688954, "grad_norm": 55.75, "loss": 3.669754, "lr_stage": "cosine_decay", "mfu": 57.682642, "stage_name": "permutation_cycle", "status": "training", "step": 655, "tokens_per_sec_per_gpu": 100415.688954, "vram_gb": 20.709735} {"adam_lr": 8.498316702566828e-05, "dt_ms": 162.760973, "global_tokens_per_sec": 100662.951937, "grad_norm": 26.75, "loss": 3.635858, "lr_stage": "cosine_decay", "mfu": 57.824679, "stage_name": "permutation_cycle", "status": "training", "step": 656, "tokens_per_sec_per_gpu": 100662.951937, "vram_gb": 20.709735} {"adam_lr": 8.49207642715503e-05, "dt_ms": 162.733078, "global_tokens_per_sec": 100680.207129, "grad_norm": 46.5, "loss": 3.300199, "lr_stage": "cosine_decay", "mfu": 57.834591, "stage_name": "permutation_cycle", "status": "training", "step": 657, "tokens_per_sec_per_gpu": 100680.207129, "vram_gb": 20.709735} {"adam_lr": 8.485825514272824e-05, "dt_ms": 163.939714, "global_tokens_per_sec": 99939.176159, "grad_norm": 21.625, "loss": 3.61411, "lr_stage": "cosine_decay", "mfu": 57.408915, "stage_name": "permutation_cycle", "status": "training", "step": 658, "tokens_per_sec_per_gpu": 99939.176159, "vram_gb": 20.709735} {"adam_lr": 8.479563982961571e-05, "dt_ms": 163.541794, "global_tokens_per_sec": 100182.342489, "grad_norm": 31.375, "loss": 3.117549, "lr_stage": "cosine_decay", "mfu": 57.548599, "stage_name": "permutation_cycle", "status": "training", "step": 659, "tokens_per_sec_per_gpu": 100182.342489, "vram_gb": 20.709735} {"adam_lr": 8.473291852294987e-05, "dt_ms": 163.814068, "global_tokens_per_sec": 100015.83024, "grad_norm": 34.75, "loss": 3.029271, "lr_stage": "cosine_decay", "mfu": 57.452948, "stage_name": "permutation_cycle", "status": "training", "step": 660, "tokens_per_sec_per_gpu": 100015.83024, "vram_gb": 20.709735} {"adam_lr": 8.467009141379065e-05, "dt_ms": 162.321568, "global_tokens_per_sec": 100935.447142, "grad_norm": 30.625, "loss": 2.785961, "lr_stage": "cosine_decay", "mfu": 57.981211, "stage_name": "permutation_cycle", "status": "training", "step": 661, "tokens_per_sec_per_gpu": 100935.447142, "vram_gb": 20.709735} {"adam_lr": 8.460715869352035e-05, "dt_ms": 163.490057, "global_tokens_per_sec": 100214.045438, "grad_norm": 25.625, "loss": 2.546681, "lr_stage": "cosine_decay", "mfu": 57.56681, "stage_name": "permutation_cycle", "status": "training", "step": 662, "tokens_per_sec_per_gpu": 100214.045438, "vram_gb": 20.709735} {"adam_lr": 8.454412055384292e-05, "dt_ms": 163.319349, "global_tokens_per_sec": 100318.793035, "grad_norm": 34.0, "loss": 3.268071, "lr_stage": "cosine_decay", "mfu": 57.626981, "stage_name": "permutation_cycle", "status": "training", "step": 663, "tokens_per_sec_per_gpu": 100318.793035, "vram_gb": 20.709735} {"adam_lr": 8.44809771867835e-05, "dt_ms": 161.601305, "global_tokens_per_sec": 101385.319872, "grad_norm": 37.75, "loss": 3.942773, "lr_stage": "cosine_decay", "mfu": 58.239635, "stage_name": "permutation_cycle", "status": "training", "step": 664, "tokens_per_sec_per_gpu": 101385.319872, "vram_gb": 20.709735} {"adam_lr": 8.44177287846877e-05, "dt_ms": 163.603544, "global_tokens_per_sec": 100144.529732, "grad_norm": 43.0, "loss": 3.402918, "lr_stage": "cosine_decay", "mfu": 57.526878, "stage_name": "permutation_cycle", "status": "training", "step": 665, "tokens_per_sec_per_gpu": 100144.529732, "vram_gb": 20.709735} {"adam_lr": 8.435437554022115e-05, "dt_ms": 162.831068, "global_tokens_per_sec": 100619.618831, "grad_norm": 28.875, "loss": 2.303895, "lr_stage": "cosine_decay", "mfu": 57.799787, "stage_name": "permutation_cycle", "status": "training", "step": 666, "tokens_per_sec_per_gpu": 100619.618831, "vram_gb": 20.709735} {"adam_lr": 8.429091764636883e-05, "dt_ms": 163.529158, "global_tokens_per_sec": 100190.083754, "grad_norm": 32.75, "loss": 3.708527, "lr_stage": "cosine_decay", "mfu": 57.553046, "stage_name": "permutation_cycle", "status": "training", "step": 667, "tokens_per_sec_per_gpu": 100190.083754, "vram_gb": 20.709735} {"adam_lr": 8.422735529643444e-05, "dt_ms": 163.944721, "global_tokens_per_sec": 99936.124066, "grad_norm": 28.625, "loss": 3.462687, "lr_stage": "cosine_decay", "mfu": 57.407161, "stage_name": "permutation_cycle", "status": "training", "step": 668, "tokens_per_sec_per_gpu": 99936.124066, "vram_gb": 20.709735} {"adam_lr": 8.416368868403997e-05, "dt_ms": 163.562536, "global_tokens_per_sec": 100169.637722, "grad_norm": 20.75, "loss": 3.001845, "lr_stage": "cosine_decay", "mfu": 57.541301, "stage_name": "permutation_cycle", "status": "training", "step": 669, "tokens_per_sec_per_gpu": 100169.637722, "vram_gb": 20.709735} {"adam_lr": 8.409991800312493e-05, "dt_ms": 162.656307, "global_tokens_per_sec": 100727.726333, "grad_norm": 27.375, "loss": 3.77664, "lr_stage": "cosine_decay", "mfu": 57.861888, "stage_name": "permutation_cycle", "status": "training", "step": 670, "tokens_per_sec_per_gpu": 100727.726333, "vram_gb": 20.709735} {"adam_lr": 8.40360434479459e-05, "dt_ms": 164.125204, "global_tokens_per_sec": 99826.227734, "grad_norm": 32.25, "loss": 3.828089, "lr_stage": "cosine_decay", "mfu": 57.344033, "stage_name": "permutation_cycle", "status": "training", "step": 671, "tokens_per_sec_per_gpu": 99826.227734, "vram_gb": 20.709735} {"adam_lr": 8.397206521307584e-05, "dt_ms": 164.836168, "global_tokens_per_sec": 99395.661584, "grad_norm": 38.0, "loss": 3.736099, "lr_stage": "cosine_decay", "mfu": 57.096699, "stage_name": "permutation_cycle", "status": "training", "step": 672, "tokens_per_sec_per_gpu": 99395.661584, "vram_gb": 20.709735} {"adam_lr": 8.390798349340355e-05, "dt_ms": 162.428141, "global_tokens_per_sec": 100869.220909, "grad_norm": 36.0, "loss": 2.806794, "lr_stage": "cosine_decay", "mfu": 57.943168, "stage_name": "permutation_cycle", "status": "training", "step": 673, "tokens_per_sec_per_gpu": 100869.220909, "vram_gb": 20.709735} {"adam_lr": 8.384379848413304e-05, "dt_ms": 162.196875, "global_tokens_per_sec": 101013.0438, "grad_norm": 34.75, "loss": 2.374152, "lr_stage": "cosine_decay", "mfu": 58.025786, "stage_name": "permutation_cycle", "status": "training", "step": 674, "tokens_per_sec_per_gpu": 101013.0438, "vram_gb": 20.709735} {"adam_lr": 8.377951038078302e-05, "dt_ms": 162.305593, "global_tokens_per_sec": 100945.381164, "grad_norm": 32.0, "loss": 3.157053, "lr_stage": "cosine_decay", "mfu": 57.986918, "stage_name": "permutation_cycle", "status": "training", "step": 675, "tokens_per_sec_per_gpu": 100945.381164, "vram_gb": 20.709735} {"adam_lr": 8.371511937918616e-05, "dt_ms": 162.11462, "global_tokens_per_sec": 101064.29623, "grad_norm": 29.5, "loss": 3.032331, "lr_stage": "cosine_decay", "mfu": 58.055227, "stage_name": "permutation_cycle", "status": "training", "step": 676, "tokens_per_sec_per_gpu": 101064.29623, "vram_gb": 20.709735} {"adam_lr": 8.365062567548867e-05, "dt_ms": 164.247036, "global_tokens_per_sec": 99752.180624, "grad_norm": 31.75, "loss": 3.195217, "lr_stage": "cosine_decay", "mfu": 57.301497, "stage_name": "permutation_cycle", "status": "training", "step": 677, "tokens_per_sec_per_gpu": 99752.180624, "vram_gb": 20.709735} {"adam_lr": 8.358602946614951e-05, "dt_ms": 163.019896, "global_tokens_per_sec": 100503.070158, "grad_norm": 22.25, "loss": 2.894937, "lr_stage": "cosine_decay", "mfu": 57.732837, "stage_name": "permutation_cycle", "status": "training", "step": 678, "tokens_per_sec_per_gpu": 100503.070158, "vram_gb": 20.709735} {"adam_lr": 8.352133094793995e-05, "dt_ms": 166.04352, "global_tokens_per_sec": 98672.926246, "grad_norm": 38.25, "loss": 4.512506, "lr_stage": "cosine_decay", "mfu": 56.681532, "stage_name": "permutation_cycle", "status": "training", "step": 679, "tokens_per_sec_per_gpu": 98672.926246, "vram_gb": 20.709735} {"adam_lr": 8.345653031794292e-05, "dt_ms": 163.003922, "global_tokens_per_sec": 100512.91925, "grad_norm": 40.5, "loss": 3.516194, "lr_stage": "cosine_decay", "mfu": 57.738495, "stage_name": "permutation_cycle", "status": "training", "step": 680, "tokens_per_sec_per_gpu": 100512.91925, "vram_gb": 20.709735} {"adam_lr": 8.339162777355234e-05, "dt_ms": 163.695335, "global_tokens_per_sec": 100088.374303, "grad_norm": 29.25, "loss": 3.392111, "lr_stage": "cosine_decay", "mfu": 57.49462, "stage_name": "permutation_cycle", "status": "training", "step": 681, "tokens_per_sec_per_gpu": 100088.374303, "vram_gb": 20.709735} {"adam_lr": 8.332662351247262e-05, "dt_ms": 161.791325, "global_tokens_per_sec": 101266.245511, "grad_norm": 34.25, "loss": 3.926653, "lr_stage": "cosine_decay", "mfu": 58.171234, "stage_name": "permutation_cycle", "status": "training", "step": 682, "tokens_per_sec_per_gpu": 101266.245511, "vram_gb": 20.709735} {"adam_lr": 8.326151773271804e-05, "dt_ms": 163.697004, "global_tokens_per_sec": 100087.353878, "grad_norm": 31.25, "loss": 3.789878, "lr_stage": "cosine_decay", "mfu": 57.494034, "stage_name": "permutation_cycle", "status": "training", "step": 683, "tokens_per_sec_per_gpu": 100087.353878, "vram_gb": 20.709735} {"adam_lr": 8.319631063261209e-05, "dt_ms": 164.195538, "global_tokens_per_sec": 99783.466973, "grad_norm": 28.125, "loss": 3.175634, "lr_stage": "cosine_decay", "mfu": 57.319469, "stage_name": "permutation_cycle", "status": "training", "step": 684, "tokens_per_sec_per_gpu": 99783.466973, "vram_gb": 20.709735} {"adam_lr": 8.313100241078689e-05, "dt_ms": 164.460182, "global_tokens_per_sec": 99622.898271, "grad_norm": 29.0, "loss": 2.667138, "lr_stage": "cosine_decay", "mfu": 57.227232, "stage_name": "permutation_cycle", "status": "training", "step": 685, "tokens_per_sec_per_gpu": 99622.898271, "vram_gb": 20.709735} {"adam_lr": 8.306559326618259e-05, "dt_ms": 163.099527, "global_tokens_per_sec": 100454.000482, "grad_norm": 51.5, "loss": 3.20055, "lr_stage": "cosine_decay", "mfu": 57.70465, "stage_name": "permutation_cycle", "status": "training", "step": 686, "tokens_per_sec_per_gpu": 100454.000482, "vram_gb": 20.709735} {"adam_lr": 8.300008339804685e-05, "dt_ms": 164.569378, "global_tokens_per_sec": 99556.796101, "grad_norm": 26.875, "loss": 3.749038, "lr_stage": "cosine_decay", "mfu": 57.189261, "stage_name": "permutation_cycle", "status": "training", "step": 687, "tokens_per_sec_per_gpu": 99556.796101, "vram_gb": 20.709735} {"adam_lr": 8.293447300593402e-05, "dt_ms": 160.666466, "global_tokens_per_sec": 101975.23125, "grad_norm": 42.25, "loss": 2.603676, "lr_stage": "cosine_decay", "mfu": 58.578503, "stage_name": "permutation_cycle", "status": "training", "step": 688, "tokens_per_sec_per_gpu": 101975.23125, "vram_gb": 20.709735} {"adam_lr": 8.28687622897048e-05, "dt_ms": 163.578987, "global_tokens_per_sec": 100159.563819, "grad_norm": 44.0, "loss": 3.95468, "lr_stage": "cosine_decay", "mfu": 57.535514, "stage_name": "permutation_cycle", "status": "training", "step": 689, "tokens_per_sec_per_gpu": 100159.563819, "vram_gb": 20.709735} {"adam_lr": 8.280295144952536e-05, "dt_ms": 162.878036, "global_tokens_per_sec": 100590.603572, "grad_norm": 46.5, "loss": 3.630927, "lr_stage": "cosine_decay", "mfu": 57.78312, "stage_name": "permutation_cycle", "status": "training", "step": 690, "tokens_per_sec_per_gpu": 100590.603572, "vram_gb": 20.709735} {"adam_lr": 8.273704068586698e-05, "dt_ms": 164.877415, "global_tokens_per_sec": 99370.796355, "grad_norm": 32.75, "loss": 3.653939, "lr_stage": "cosine_decay", "mfu": 57.082415, "stage_name": "permutation_cycle", "status": "training", "step": 691, "tokens_per_sec_per_gpu": 99370.796355, "vram_gb": 20.709735} {"adam_lr": 8.267103019950529e-05, "dt_ms": 163.903713, "global_tokens_per_sec": 99961.127649, "grad_norm": 29.875, "loss": 3.523358, "lr_stage": "cosine_decay", "mfu": 57.421524, "stage_name": "permutation_cycle", "status": "training", "step": 692, "tokens_per_sec_per_gpu": 99961.127649, "vram_gb": 20.709735} {"adam_lr": 8.260492019151962e-05, "dt_ms": 162.624121, "global_tokens_per_sec": 100747.662329, "grad_norm": 35.5, "loss": 3.216876, "lr_stage": "cosine_decay", "mfu": 57.87334, "stage_name": "permutation_cycle", "status": "training", "step": 693, "tokens_per_sec_per_gpu": 100747.662329, "vram_gb": 20.709735} {"adam_lr": 8.253871086329255e-05, "dt_ms": 161.991835, "global_tokens_per_sec": 101140.90032, "grad_norm": 35.5, "loss": 3.147168, "lr_stage": "cosine_decay", "mfu": 58.099231, "stage_name": "permutation_cycle", "status": "training", "step": 694, "tokens_per_sec_per_gpu": 101140.90032, "vram_gb": 20.709735} {"adam_lr": 8.247240241650918e-05, "dt_ms": 164.027214, "global_tokens_per_sec": 99885.864031, "grad_norm": 41.75, "loss": 3.63253, "lr_stage": "cosine_decay", "mfu": 57.37829, "stage_name": "permutation_cycle", "status": "training", "step": 695, "tokens_per_sec_per_gpu": 99885.864031, "vram_gb": 20.709735} {"adam_lr": 8.240599505315655e-05, "dt_ms": 165.400982, "global_tokens_per_sec": 99056.243872, "grad_norm": 37.25, "loss": 4.223212, "lr_stage": "cosine_decay", "mfu": 56.901724, "stage_name": "permutation_cycle", "status": "training", "step": 696, "tokens_per_sec_per_gpu": 99056.243872, "vram_gb": 20.709735} {"adam_lr": 8.2339488975523e-05, "dt_ms": 163.660288, "global_tokens_per_sec": 100109.808033, "grad_norm": 25.875, "loss": 2.432592, "lr_stage": "cosine_decay", "mfu": 57.506932, "stage_name": "permutation_cycle", "status": "training", "step": 697, "tokens_per_sec_per_gpu": 100109.808033, "vram_gb": 20.709735} {"adam_lr": 8.227288438619754e-05, "dt_ms": 162.899256, "global_tokens_per_sec": 100577.500642, "grad_norm": 29.375, "loss": 3.069179, "lr_stage": "cosine_decay", "mfu": 57.775593, "stage_name": "permutation_cycle", "status": "training", "step": 698, "tokens_per_sec_per_gpu": 100577.500642, "vram_gb": 20.709735} {"adam_lr": 8.220618148806933e-05, "dt_ms": 162.833691, "global_tokens_per_sec": 100617.998249, "grad_norm": 37.0, "loss": 3.183494, "lr_stage": "cosine_decay", "mfu": 57.798856, "stage_name": "permutation_cycle", "status": "training", "step": 699, "tokens_per_sec_per_gpu": 100617.998249, "vram_gb": 20.709735} {"adam_lr": 8.213938048432697e-05, "dt_ms": 162.630796, "global_tokens_per_sec": 100743.526807, "grad_norm": 42.0, "loss": 2.247973, "lr_stage": "cosine_decay", "mfu": 57.870965, "stage_name": "permutation_cycle", "status": "training", "step": 700, "tokens_per_sec_per_gpu": 100743.526807, "vram_gb": 20.709735} {"adam_lr": 8.207248157845791e-05, "dt_ms": 161.720753, "global_tokens_per_sec": 101310.436198, "grad_norm": 33.0, "loss": 3.393145, "lr_stage": "cosine_decay", "mfu": 58.196619, "stage_name": "permutation_cycle", "status": "training", "step": 701, "tokens_per_sec_per_gpu": 101310.436198, "vram_gb": 20.709735} {"adam_lr": 8.200548497424778e-05, "dt_ms": 163.139343, "global_tokens_per_sec": 100429.483608, "grad_norm": 30.25, "loss": 2.877368, "lr_stage": "cosine_decay", "mfu": 57.690566, "stage_name": "permutation_cycle", "status": "training", "step": 702, "tokens_per_sec_per_gpu": 100429.483608, "vram_gb": 20.709735} {"adam_lr": 8.193839087577989e-05, "dt_ms": 162.382841, "global_tokens_per_sec": 100897.36014, "grad_norm": 34.5, "loss": 3.453747, "lr_stage": "cosine_decay", "mfu": 57.959332, "stage_name": "permutation_cycle", "status": "training", "step": 703, "tokens_per_sec_per_gpu": 100897.36014, "vram_gb": 20.709735} {"adam_lr": 8.18711994874345e-05, "dt_ms": 163.724661, "global_tokens_per_sec": 100070.447009, "grad_norm": 27.5, "loss": 2.76762, "lr_stage": "cosine_decay", "mfu": 57.484322, "stage_name": "permutation_cycle", "status": "training", "step": 704, "tokens_per_sec_per_gpu": 100070.447009, "vram_gb": 20.709735} {"adam_lr": 8.18039110138882e-05, "dt_ms": 161.854029, "global_tokens_per_sec": 101227.013819, "grad_norm": 32.5, "loss": 3.659299, "lr_stage": "cosine_decay", "mfu": 58.148698, "stage_name": "permutation_cycle", "status": "training", "step": 705, "tokens_per_sec_per_gpu": 101227.013819, "vram_gb": 20.709735} {"adam_lr": 8.173652566011338e-05, "dt_ms": 163.427114, "global_tokens_per_sec": 100252.641989, "grad_norm": 25.875, "loss": 3.725271, "lr_stage": "cosine_decay", "mfu": 57.588982, "stage_name": "permutation_cycle", "status": "training", "step": 706, "tokens_per_sec_per_gpu": 100252.641989, "vram_gb": 20.709735} {"adam_lr": 8.16690436313775e-05, "dt_ms": 163.103819, "global_tokens_per_sec": 100451.35737, "grad_norm": 34.5, "loss": 5.169447, "lr_stage": "cosine_decay", "mfu": 57.703131, "stage_name": "permutation_cycle", "status": "training", "step": 707, "tokens_per_sec_per_gpu": 100451.35737, "vram_gb": 20.709735} {"adam_lr": 8.160146513324254e-05, "dt_ms": 163.352251, "global_tokens_per_sec": 100298.587221, "grad_norm": 33.0, "loss": 3.309865, "lr_stage": "cosine_decay", "mfu": 57.615374, "stage_name": "permutation_cycle", "status": "training", "step": 708, "tokens_per_sec_per_gpu": 100298.587221, "vram_gb": 20.709735} {"adam_lr": 8.153379037156432e-05, "dt_ms": 164.197445, "global_tokens_per_sec": 99782.30787, "grad_norm": 25.75, "loss": 3.107612, "lr_stage": "cosine_decay", "mfu": 57.318803, "stage_name": "permutation_cycle", "status": "training", "step": 709, "tokens_per_sec_per_gpu": 99782.30787, "vram_gb": 20.709735} {"adam_lr": 8.146601955249188e-05, "dt_ms": 161.147356, "global_tokens_per_sec": 101670.920351, "grad_norm": 26.375, "loss": 3.563714, "lr_stage": "cosine_decay", "mfu": 58.403695, "stage_name": "permutation_cycle", "status": "training", "step": 710, "tokens_per_sec_per_gpu": 101670.920351, "vram_gb": 20.709735} {"adam_lr": 8.139815288246691e-05, "dt_ms": 164.79063, "global_tokens_per_sec": 99423.128403, "grad_norm": 34.75, "loss": 2.948048, "lr_stage": "cosine_decay", "mfu": 57.112477, "stage_name": "permutation_cycle", "status": "training", "step": 711, "tokens_per_sec_per_gpu": 99423.128403, "vram_gb": 20.709735} {"adam_lr": 8.133019056822304e-05, "dt_ms": 159.810066, "global_tokens_per_sec": 102521.702088, "grad_norm": 23.875, "loss": 2.117476, "lr_stage": "cosine_decay", "mfu": 58.892417, "stage_name": "permutation_cycle", "status": "training", "step": 712, "tokens_per_sec_per_gpu": 102521.702088, "vram_gb": 20.709735} {"adam_lr": 8.126213281678528e-05, "dt_ms": 166.945696, "global_tokens_per_sec": 98139.696947, "grad_norm": 25.25, "loss": 3.741991, "lr_stage": "cosine_decay", "mfu": 56.375224, "stage_name": "permutation_cycle", "status": "training", "step": 713, "tokens_per_sec_per_gpu": 98139.696947, "vram_gb": 20.709735} {"adam_lr": 8.119397983546932e-05, "dt_ms": 164.524794, "global_tokens_per_sec": 99583.774816, "grad_norm": 35.0, "loss": 3.595187, "lr_stage": "cosine_decay", "mfu": 57.204758, "stage_name": "permutation_cycle", "status": "training", "step": 714, "tokens_per_sec_per_gpu": 99583.774816, "vram_gb": 20.709735} {"adam_lr": 8.112573183188099e-05, "dt_ms": 163.245678, "global_tokens_per_sec": 100364.066026, "grad_norm": 38.0, "loss": 2.801862, "lr_stage": "cosine_decay", "mfu": 57.652988, "stage_name": "permutation_cycle", "status": "training", "step": 715, "tokens_per_sec_per_gpu": 100364.066026, "vram_gb": 20.709735} {"adam_lr": 8.105738901391552e-05, "dt_ms": 162.262201, "global_tokens_per_sec": 100972.375993, "grad_norm": 29.875, "loss": 3.521986, "lr_stage": "cosine_decay", "mfu": 58.002424, "stage_name": "permutation_cycle", "status": "training", "step": 716, "tokens_per_sec_per_gpu": 100972.375993, "vram_gb": 20.709735} {"adam_lr": 8.098895158975702e-05, "dt_ms": 163.615704, "global_tokens_per_sec": 100137.087341, "grad_norm": 35.75, "loss": 3.515436, "lr_stage": "cosine_decay", "mfu": 57.522602, "stage_name": "permutation_cycle", "status": "training", "step": 717, "tokens_per_sec_per_gpu": 100137.087341, "vram_gb": 20.709735} {"adam_lr": 8.09204197678777e-05, "dt_ms": 159.738779, "global_tokens_per_sec": 102567.454788, "grad_norm": 34.0, "loss": 2.529663, "lr_stage": "cosine_decay", "mfu": 58.918699, "stage_name": "permutation_cycle", "status": "training", "step": 718, "tokens_per_sec_per_gpu": 102567.454788, "vram_gb": 20.709735} {"adam_lr": 8.085179375703744e-05, "dt_ms": 163.623095, "global_tokens_per_sec": 100132.564075, "grad_norm": 46.25, "loss": 4.343877, "lr_stage": "cosine_decay", "mfu": 57.520004, "stage_name": "permutation_cycle", "status": "training", "step": 719, "tokens_per_sec_per_gpu": 100132.564075, "vram_gb": 20.709735} {"adam_lr": 8.07830737662829e-05, "dt_ms": 165.578604, "global_tokens_per_sec": 98949.982845, "grad_norm": 36.75, "loss": 3.881258, "lr_stage": "cosine_decay", "mfu": 56.840684, "stage_name": "permutation_cycle", "status": "training", "step": 720, "tokens_per_sec_per_gpu": 98949.982845, "vram_gb": 20.709735} {"adam_lr": 8.071426000494716e-05, "dt_ms": 163.908958, "global_tokens_per_sec": 99957.928819, "grad_norm": 23.75, "loss": 2.535354, "lr_stage": "cosine_decay", "mfu": 57.419687, "stage_name": "permutation_cycle", "status": "training", "step": 721, "tokens_per_sec_per_gpu": 99957.928819, "vram_gb": 20.709735} {"adam_lr": 8.064535268264883e-05, "dt_ms": 163.921118, "global_tokens_per_sec": 99950.514135, "grad_norm": 41.5, "loss": 3.452047, "lr_stage": "cosine_decay", "mfu": 57.415428, "stage_name": "permutation_cycle", "status": "training", "step": 722, "tokens_per_sec_per_gpu": 99950.514135, "vram_gb": 20.709735} {"adam_lr": 8.057635200929157e-05, "dt_ms": 163.542747, "global_tokens_per_sec": 100181.758291, "grad_norm": 28.375, "loss": 3.415551, "lr_stage": "cosine_decay", "mfu": 57.548263, "stage_name": "permutation_cycle", "status": "training", "step": 723, "tokens_per_sec_per_gpu": 100181.758291, "vram_gb": 20.709735} {"adam_lr": 8.05072581950634e-05, "dt_ms": 162.62126, "global_tokens_per_sec": 100749.434799, "grad_norm": 27.5, "loss": 3.746792, "lr_stage": "cosine_decay", "mfu": 57.874358, "stage_name": "permutation_cycle", "status": "training", "step": 724, "tokens_per_sec_per_gpu": 100749.434799, "vram_gb": 20.709735} {"adam_lr": 8.043807145043604e-05, "dt_ms": 163.406134, "global_tokens_per_sec": 100265.514114, "grad_norm": 34.0, "loss": 3.720656, "lr_stage": "cosine_decay", "mfu": 57.596376, "stage_name": "permutation_cycle", "status": "training", "step": 725, "tokens_per_sec_per_gpu": 100265.514114, "vram_gb": 20.709735} {"adam_lr": 8.036879198616434e-05, "dt_ms": 164.372206, "global_tokens_per_sec": 99676.219144, "grad_norm": 35.0, "loss": 2.809951, "lr_stage": "cosine_decay", "mfu": 57.257862, "stage_name": "permutation_cycle", "status": "training", "step": 726, "tokens_per_sec_per_gpu": 99676.219144, "vram_gb": 20.709735} {"adam_lr": 8.029942001328556e-05, "dt_ms": 162.898064, "global_tokens_per_sec": 100578.236671, "grad_norm": 34.75, "loss": 2.931186, "lr_stage": "cosine_decay", "mfu": 57.776016, "stage_name": "permutation_cycle", "status": "training", "step": 727, "tokens_per_sec_per_gpu": 100578.236671, "vram_gb": 20.709735} {"adam_lr": 8.022995574311876e-05, "dt_ms": 163.039684, "global_tokens_per_sec": 100490.871721, "grad_norm": 34.5, "loss": 2.85288, "lr_stage": "cosine_decay", "mfu": 57.72583, "stage_name": "permutation_cycle", "status": "training", "step": 728, "tokens_per_sec_per_gpu": 100490.871721, "vram_gb": 20.709735} {"adam_lr": 8.016039938726412e-05, "dt_ms": 163.677454, "global_tokens_per_sec": 100099.308733, "grad_norm": 28.625, "loss": 2.761215, "lr_stage": "cosine_decay", "mfu": 57.500901, "stage_name": "permutation_cycle", "status": "training", "step": 729, "tokens_per_sec_per_gpu": 100099.308733, "vram_gb": 20.709735} {"adam_lr": 8.009075115760243e-05, "dt_ms": 162.641048, "global_tokens_per_sec": 100737.176488, "grad_norm": 33.0, "loss": 3.503186, "lr_stage": "cosine_decay", "mfu": 57.867317, "stage_name": "permutation_cycle", "status": "training", "step": 730, "tokens_per_sec_per_gpu": 100737.176488, "vram_gb": 20.709735} {"adam_lr": 8.002101126629421e-05, "dt_ms": 164.033651, "global_tokens_per_sec": 99881.944131, "grad_norm": 26.25, "loss": 2.817944, "lr_stage": "cosine_decay", "mfu": 57.376038, "stage_name": "permutation_cycle", "status": "training", "step": 731, "tokens_per_sec_per_gpu": 99881.944131, "vram_gb": 20.709735} {"adam_lr": 7.99511799257793e-05, "dt_ms": 162.548065, "global_tokens_per_sec": 100794.801718, "grad_norm": 44.0, "loss": 4.056913, "lr_stage": "cosine_decay", "mfu": 57.900419, "stage_name": "permutation_cycle", "status": "training", "step": 732, "tokens_per_sec_per_gpu": 100794.801718, "vram_gb": 20.709735} {"adam_lr": 7.988125734877607e-05, "dt_ms": 162.835121, "global_tokens_per_sec": 100617.114317, "grad_norm": 36.0, "loss": 2.293797, "lr_stage": "cosine_decay", "mfu": 57.798348, "stage_name": "permutation_cycle", "status": "training", "step": 733, "tokens_per_sec_per_gpu": 100617.114317, "vram_gb": 20.709735} {"adam_lr": 7.98112437482808e-05, "dt_ms": 162.692785, "global_tokens_per_sec": 100705.14174, "grad_norm": 50.5, "loss": 3.153488, "lr_stage": "cosine_decay", "mfu": 57.848915, "stage_name": "permutation_cycle", "status": "training", "step": 734, "tokens_per_sec_per_gpu": 100705.14174, "vram_gb": 20.709735} {"adam_lr": 7.974113933756707e-05, "dt_ms": 164.493561, "global_tokens_per_sec": 99602.683055, "grad_norm": 37.0, "loss": 3.133527, "lr_stage": "cosine_decay", "mfu": 57.21562, "stage_name": "permutation_cycle", "status": "training", "step": 735, "tokens_per_sec_per_gpu": 99602.683055, "vram_gb": 20.709735} {"adam_lr": 7.967094433018508e-05, "dt_ms": 164.085627, "global_tokens_per_sec": 99850.305839, "grad_norm": 42.25, "loss": 3.384892, "lr_stage": "cosine_decay", "mfu": 57.357864, "stage_name": "permutation_cycle", "status": "training", "step": 736, "tokens_per_sec_per_gpu": 99850.305839, "vram_gb": 20.709735} {"adam_lr": 7.960065893996098e-05, "dt_ms": 162.222862, "global_tokens_per_sec": 100996.861807, "grad_norm": 25.875, "loss": 2.948252, "lr_stage": "cosine_decay", "mfu": 58.01649, "stage_name": "permutation_cycle", "status": "training", "step": 737, "tokens_per_sec_per_gpu": 100996.861807, "vram_gb": 20.709735} {"adam_lr": 7.953028338099627e-05, "dt_ms": 163.058996, "global_tokens_per_sec": 100478.970077, "grad_norm": 35.25, "loss": 4.098913, "lr_stage": "cosine_decay", "mfu": 57.718993, "stage_name": "permutation_cycle", "status": "training", "step": 738, "tokens_per_sec_per_gpu": 100478.970077, "vram_gb": 20.709735} {"adam_lr": 7.94598178676671e-05, "dt_ms": 165.261507, "global_tokens_per_sec": 99139.843839, "grad_norm": 48.0, "loss": 3.34873, "lr_stage": "cosine_decay", "mfu": 56.949747, "stage_name": "permutation_cycle", "status": "training", "step": 739, "tokens_per_sec_per_gpu": 99139.843839, "vram_gb": 20.709735} {"adam_lr": 7.938926261462366e-05, "dt_ms": 162.409782, "global_tokens_per_sec": 100880.62281, "grad_norm": 37.25, "loss": 3.500453, "lr_stage": "cosine_decay", "mfu": 57.949718, "stage_name": "permutation_cycle", "status": "training", "step": 740, "tokens_per_sec_per_gpu": 100880.62281, "vram_gb": 20.709735} {"adam_lr": 7.931861783678946e-05, "dt_ms": 162.2684, "global_tokens_per_sec": 100968.518705, "grad_norm": 32.75, "loss": 3.700943, "lr_stage": "cosine_decay", "mfu": 58.000209, "stage_name": "permutation_cycle", "status": "training", "step": 741, "tokens_per_sec_per_gpu": 100968.518705, "vram_gb": 20.709735} {"adam_lr": 7.924788374936078e-05, "dt_ms": 162.630796, "global_tokens_per_sec": 100743.526807, "grad_norm": 26.375, "loss": 3.724944, "lr_stage": "cosine_decay", "mfu": 57.870965, "stage_name": "permutation_cycle", "status": "training", "step": 742, "tokens_per_sec_per_gpu": 100743.526807, "vram_gb": 20.709735} {"adam_lr": 7.917706056780587e-05, "dt_ms": 162.822723, "global_tokens_per_sec": 100624.775578, "grad_norm": 48.0, "loss": 2.669682, "lr_stage": "cosine_decay", "mfu": 57.802749, "stage_name": "permutation_cycle", "status": "training", "step": 743, "tokens_per_sec_per_gpu": 100624.775578, "vram_gb": 20.709735} {"adam_lr": 7.910614850786448e-05, "dt_ms": 162.662745, "global_tokens_per_sec": 100723.74008, "grad_norm": 35.25, "loss": 3.619074, "lr_stage": "cosine_decay", "mfu": 57.859598, "stage_name": "permutation_cycle", "status": "training", "step": 744, "tokens_per_sec_per_gpu": 100723.74008, "vram_gb": 20.709735} {"adam_lr": 7.903514778554699e-05, "dt_ms": 162.75692, "global_tokens_per_sec": 100665.458734, "grad_norm": 35.25, "loss": 2.97367, "lr_stage": "cosine_decay", "mfu": 57.826119, "stage_name": "permutation_cycle", "status": "training", "step": 745, "tokens_per_sec_per_gpu": 100665.458734, "vram_gb": 20.709735} {"adam_lr": 7.896405861713394e-05, "dt_ms": 164.543629, "global_tokens_per_sec": 99572.375607, "grad_norm": 22.625, "loss": 3.114786, "lr_stage": "cosine_decay", "mfu": 57.19821, "stage_name": "permutation_cycle", "status": "training", "step": 746, "tokens_per_sec_per_gpu": 99572.375607, "vram_gb": 20.709735} {"adam_lr": 7.889288121917527e-05, "dt_ms": 163.19418, "global_tokens_per_sec": 100395.737438, "grad_norm": 27.75, "loss": 2.032683, "lr_stage": "cosine_decay", "mfu": 57.671181, "stage_name": "permutation_cycle", "status": "training", "step": 747, "tokens_per_sec_per_gpu": 100395.737438, "vram_gb": 20.709735} {"adam_lr": 7.882161580848967e-05, "dt_ms": 162.897825, "global_tokens_per_sec": 100578.383878, "grad_norm": 40.75, "loss": 4.166151, "lr_stage": "cosine_decay", "mfu": 57.7761, "stage_name": "permutation_cycle", "status": "training", "step": 748, "tokens_per_sec_per_gpu": 100578.383878, "vram_gb": 20.709735} {"adam_lr": 7.875026260216393e-05, "dt_ms": 163.711548, "global_tokens_per_sec": 100078.462485, "grad_norm": 43.0, "loss": 3.437664, "lr_stage": "cosine_decay", "mfu": 57.488926, "stage_name": "permutation_cycle", "status": "training", "step": 749, "tokens_per_sec_per_gpu": 100078.462485, "vram_gb": 20.709735} {"adam_lr": 7.86788218175523e-05, "dt_ms": 164.739609, "global_tokens_per_sec": 99453.920784, "grad_norm": 29.625, "loss": 2.310322, "lr_stage": "cosine_decay", "mfu": 57.130165, "stage_name": "permutation_cycle", "status": "training", "step": 750, "tokens_per_sec_per_gpu": 99453.920784, "vram_gb": 20.709735} {"adam_lr": 7.860729367227582e-05, "dt_ms": 162.997246, "global_tokens_per_sec": 100517.03586, "grad_norm": 28.875, "loss": 2.815419, "lr_stage": "cosine_decay", "mfu": 57.74086, "stage_name": "permutation_cycle", "status": "training", "step": 751, "tokens_per_sec_per_gpu": 100517.03586, "vram_gb": 20.709735} {"adam_lr": 7.85356783842216e-05, "dt_ms": 162.237167, "global_tokens_per_sec": 100987.956501, "grad_norm": 34.25, "loss": 3.671112, "lr_stage": "cosine_decay", "mfu": 58.011375, "stage_name": "permutation_cycle", "status": "training", "step": 752, "tokens_per_sec_per_gpu": 100987.956501, "vram_gb": 20.709735} {"adam_lr": 7.846397617154222e-05, "dt_ms": 163.541555, "global_tokens_per_sec": 100182.488539, "grad_norm": 38.75, "loss": 3.743966, "lr_stage": "cosine_decay", "mfu": 57.548683, "stage_name": "permutation_cycle", "status": "training", "step": 753, "tokens_per_sec_per_gpu": 100182.488539, "vram_gb": 20.709735} {"adam_lr": 7.839218725265506e-05, "dt_ms": 161.854982, "global_tokens_per_sec": 101226.417374, "grad_norm": 34.0, "loss": 3.716534, "lr_stage": "cosine_decay", "mfu": 58.148356, "stage_name": "permutation_cycle", "status": "training", "step": 754, "tokens_per_sec_per_gpu": 101226.417374, "vram_gb": 20.709735} {"adam_lr": 7.832031184624164e-05, "dt_ms": 161.09848, "global_tokens_per_sec": 101701.76638, "grad_norm": 42.0, "loss": 2.809078, "lr_stage": "cosine_decay", "mfu": 58.421414, "stage_name": "permutation_cycle", "status": "training", "step": 755, "tokens_per_sec_per_gpu": 101701.76638, "vram_gb": 20.709735} {"adam_lr": 7.82483501712469e-05, "dt_ms": 165.08913, "global_tokens_per_sec": 99243.359997, "grad_norm": 30.875, "loss": 3.380053, "lr_stage": "cosine_decay", "mfu": 57.009211, "stage_name": "permutation_cycle", "status": "training", "step": 756, "tokens_per_sec_per_gpu": 99243.359997, "vram_gb": 20.709735} {"adam_lr": 7.817630244687857e-05, "dt_ms": 163.543224, "global_tokens_per_sec": 100181.466194, "grad_norm": 30.375, "loss": 3.738213, "lr_stage": "cosine_decay", "mfu": 57.548095, "stage_name": "permutation_cycle", "status": "training", "step": 757, "tokens_per_sec_per_gpu": 100181.466194, "vram_gb": 20.709735} {"adam_lr": 7.810416889260653e-05, "dt_ms": 163.390875, "global_tokens_per_sec": 100274.877736, "grad_norm": 27.125, "loss": 3.795438, "lr_stage": "cosine_decay", "mfu": 57.601755, "stage_name": "permutation_cycle", "status": "training", "step": 758, "tokens_per_sec_per_gpu": 100274.877736, "vram_gb": 20.709735} {"adam_lr": 7.80319497281621e-05, "dt_ms": 162.832499, "global_tokens_per_sec": 100618.734871, "grad_norm": 26.25, "loss": 3.328201, "lr_stage": "cosine_decay", "mfu": 57.799279, "stage_name": "permutation_cycle", "status": "training", "step": 759, "tokens_per_sec_per_gpu": 100618.734871, "vram_gb": 20.709735} {"adam_lr": 7.795964517353735e-05, "dt_ms": 162.633657, "global_tokens_per_sec": 100741.754544, "grad_norm": 39.75, "loss": 2.925958, "lr_stage": "cosine_decay", "mfu": 57.869947, "stage_name": "permutation_cycle", "status": "training", "step": 760, "tokens_per_sec_per_gpu": 100741.754544, "vram_gb": 20.709735} {"adam_lr": 7.788725544898452e-05, "dt_ms": 162.291527, "global_tokens_per_sec": 100954.130654, "grad_norm": 22.0, "loss": 3.062125, "lr_stage": "cosine_decay", "mfu": 57.991944, "stage_name": "permutation_cycle", "status": "training", "step": 761, "tokens_per_sec_per_gpu": 100954.130654, "vram_gb": 20.709735} {"adam_lr": 7.781478077501525e-05, "dt_ms": 165.467024, "global_tokens_per_sec": 99016.708096, "grad_norm": 28.375, "loss": 3.523687, "lr_stage": "cosine_decay", "mfu": 56.879013, "stage_name": "permutation_cycle", "status": "training", "step": 762, "tokens_per_sec_per_gpu": 99016.708096, "vram_gb": 20.709735} {"adam_lr": 7.774222137239997e-05, "dt_ms": 164.744854, "global_tokens_per_sec": 99450.754332, "grad_norm": 24.375, "loss": 3.482366, "lr_stage": "cosine_decay", "mfu": 57.128346, "stage_name": "permutation_cycle", "status": "training", "step": 763, "tokens_per_sec_per_gpu": 99450.754332, "vram_gb": 20.709735} {"adam_lr": 7.766957746216721e-05, "dt_ms": 163.269281, "global_tokens_per_sec": 100349.556639, "grad_norm": 27.0, "loss": 2.982621, "lr_stage": "cosine_decay", "mfu": 57.644653, "stage_name": "permutation_cycle", "status": "training", "step": 764, "tokens_per_sec_per_gpu": 100349.556639, "vram_gb": 20.709735} {"adam_lr": 7.75968492656029e-05, "dt_ms": 164.428711, "global_tokens_per_sec": 99641.96585, "grad_norm": 35.25, "loss": 2.851892, "lr_stage": "cosine_decay", "mfu": 57.238186, "stage_name": "permutation_cycle", "status": "training", "step": 765, "tokens_per_sec_per_gpu": 99641.96585, "vram_gb": 20.709735} {"adam_lr": 7.752403700424979e-05, "dt_ms": 162.199974, "global_tokens_per_sec": 101011.113565, "grad_norm": 42.5, "loss": 2.855656, "lr_stage": "cosine_decay", "mfu": 58.024677, "stage_name": "permutation_cycle", "status": "training", "step": 766, "tokens_per_sec_per_gpu": 101011.113565, "vram_gb": 20.709735} {"adam_lr": 7.74511408999066e-05, "dt_ms": 162.622452, "global_tokens_per_sec": 100748.696262, "grad_norm": 46.0, "loss": 3.73027, "lr_stage": "cosine_decay", "mfu": 57.873934, "stage_name": "permutation_cycle", "status": "training", "step": 767, "tokens_per_sec_per_gpu": 100748.696262, "vram_gb": 20.709735} {"adam_lr": 7.737816117462752e-05, "dt_ms": 162.912607, "global_tokens_per_sec": 100569.257851, "grad_norm": 28.5, "loss": 2.83836, "lr_stage": "cosine_decay", "mfu": 57.770858, "stage_name": "permutation_cycle", "status": "training", "step": 768, "tokens_per_sec_per_gpu": 100569.257851, "vram_gb": 20.709735} {"adam_lr": 7.730509805072145e-05, "dt_ms": 163.339376, "global_tokens_per_sec": 100306.492875, "grad_norm": 47.5, "loss": 4.71149, "lr_stage": "cosine_decay", "mfu": 57.619916, "stage_name": "permutation_cycle", "status": "training", "step": 769, "tokens_per_sec_per_gpu": 100306.492875, "vram_gb": 20.709735} {"adam_lr": 7.723195175075136e-05, "dt_ms": 164.814472, "global_tokens_per_sec": 99408.745976, "grad_norm": 22.75, "loss": 3.351999, "lr_stage": "cosine_decay", "mfu": 57.104215, "stage_name": "permutation_cycle", "status": "training", "step": 770, "tokens_per_sec_per_gpu": 99408.745976, "vram_gb": 20.709735} {"adam_lr": 7.715872249753353e-05, "dt_ms": 163.858652, "global_tokens_per_sec": 99988.616948, "grad_norm": 29.25, "loss": 2.999654, "lr_stage": "cosine_decay", "mfu": 57.437315, "stage_name": "permutation_cycle", "status": "training", "step": 771, "tokens_per_sec_per_gpu": 99988.616948, "vram_gb": 20.709735} {"adam_lr": 7.7085410514137e-05, "dt_ms": 164.231062, "global_tokens_per_sec": 99761.883087, "grad_norm": 28.0, "loss": 2.857139, "lr_stage": "cosine_decay", "mfu": 57.307071, "stage_name": "permutation_cycle", "status": "training", "step": 772, "tokens_per_sec_per_gpu": 99761.883087, "vram_gb": 20.709735} {"adam_lr": 7.701201602388276e-05, "dt_ms": 161.934614, "global_tokens_per_sec": 101176.638996, "grad_norm": 38.5, "loss": 2.303645, "lr_stage": "cosine_decay", "mfu": 58.119761, "stage_name": "permutation_cycle", "status": "training", "step": 773, "tokens_per_sec_per_gpu": 101176.638996, "vram_gb": 20.709735} {"adam_lr": 7.693853925034315e-05, "dt_ms": 162.50515, "global_tokens_per_sec": 100821.42022, "grad_norm": 30.875, "loss": 3.045505, "lr_stage": "cosine_decay", "mfu": 57.91571, "stage_name": "permutation_cycle", "status": "training", "step": 774, "tokens_per_sec_per_gpu": 100821.42022, "vram_gb": 20.709735} {"adam_lr": 7.68649804173412e-05, "dt_ms": 163.181782, "global_tokens_per_sec": 100403.365023, "grad_norm": 35.0, "loss": 3.040667, "lr_stage": "cosine_decay", "mfu": 57.675563, "stage_name": "permutation_cycle", "status": "training", "step": 775, "tokens_per_sec_per_gpu": 100403.365023, "vram_gb": 20.709735} {"adam_lr": 7.679133974894983e-05, "dt_ms": 164.798737, "global_tokens_per_sec": 99418.237911, "grad_norm": 20.875, "loss": 3.272198, "lr_stage": "cosine_decay", "mfu": 57.109668, "stage_name": "permutation_cycle", "status": "training", "step": 776, "tokens_per_sec_per_gpu": 99418.237911, "vram_gb": 20.709735} {"adam_lr": 7.671761746949133e-05, "dt_ms": 164.086103, "global_tokens_per_sec": 99850.015672, "grad_norm": 33.5, "loss": 3.710345, "lr_stage": "cosine_decay", "mfu": 57.357697, "stage_name": "permutation_cycle", "status": "training", "step": 777, "tokens_per_sec_per_gpu": 99850.015672, "vram_gb": 20.709735} {"adam_lr": 7.66438138035365e-05, "dt_ms": 163.146019, "global_tokens_per_sec": 100425.374166, "grad_norm": 21.25, "loss": 2.649667, "lr_stage": "cosine_decay", "mfu": 57.688206, "stage_name": "permutation_cycle", "status": "training", "step": 778, "tokens_per_sec_per_gpu": 100425.374166, "vram_gb": 20.709735} {"adam_lr": 7.656992897590414e-05, "dt_ms": 162.899256, "global_tokens_per_sec": 100577.500642, "grad_norm": 40.75, "loss": 3.567571, "lr_stage": "cosine_decay", "mfu": 57.775593, "stage_name": "permutation_cycle", "status": "training", "step": 779, "tokens_per_sec_per_gpu": 100577.500642, "vram_gb": 20.709735} {"adam_lr": 7.649596321166024e-05, "dt_ms": 163.061142, "global_tokens_per_sec": 100477.647846, "grad_norm": 35.75, "loss": 3.114933, "lr_stage": "cosine_decay", "mfu": 57.718234, "stage_name": "permutation_cycle", "status": "training", "step": 780, "tokens_per_sec_per_gpu": 100477.647846, "vram_gb": 20.709735} {"adam_lr": 7.642191673611736e-05, "dt_ms": 162.760496, "global_tokens_per_sec": 100663.246848, "grad_norm": 40.0, "loss": 3.678355, "lr_stage": "cosine_decay", "mfu": 57.824849, "stage_name": "permutation_cycle", "status": "training", "step": 781, "tokens_per_sec_per_gpu": 100663.246848, "vram_gb": 20.709735} {"adam_lr": 7.634778977483389e-05, "dt_ms": 164.666176, "global_tokens_per_sec": 99498.272284, "grad_norm": 32.5, "loss": 3.574264, "lr_stage": "cosine_decay", "mfu": 57.155642, "stage_name": "permutation_cycle", "status": "training", "step": 782, "tokens_per_sec_per_gpu": 99498.272284, "vram_gb": 20.709735} {"adam_lr": 7.627358255361339e-05, "dt_ms": 162.3981, "global_tokens_per_sec": 100887.879908, "grad_norm": 40.0, "loss": 3.373788, "lr_stage": "cosine_decay", "mfu": 57.953887, "stage_name": "permutation_cycle", "status": "training", "step": 783, "tokens_per_sec_per_gpu": 100887.879908, "vram_gb": 20.709735} {"adam_lr": 7.619929529850397e-05, "dt_ms": 164.480925, "global_tokens_per_sec": 99610.334993, "grad_norm": 31.875, "loss": 3.39443, "lr_stage": "cosine_decay", "mfu": 57.220016, "stage_name": "permutation_cycle", "status": "training", "step": 784, "tokens_per_sec_per_gpu": 99610.334993, "vram_gb": 20.709735} {"adam_lr": 7.612492823579745e-05, "dt_ms": 163.285494, "global_tokens_per_sec": 100339.593026, "grad_norm": 34.5, "loss": 3.190864, "lr_stage": "cosine_decay", "mfu": 57.63893, "stage_name": "permutation_cycle", "status": "training", "step": 785, "tokens_per_sec_per_gpu": 100339.593026, "vram_gb": 20.709735} {"adam_lr": 7.605048159202883e-05, "dt_ms": 162.576914, "global_tokens_per_sec": 100776.916068, "grad_norm": 38.5, "loss": 3.277218, "lr_stage": "cosine_decay", "mfu": 57.890145, "stage_name": "permutation_cycle", "status": "training", "step": 786, "tokens_per_sec_per_gpu": 100776.916068, "vram_gb": 20.709735} {"adam_lr": 7.597595559397548e-05, "dt_ms": 162.128925, "global_tokens_per_sec": 101055.379028, "grad_norm": 34.75, "loss": 3.174566, "lr_stage": "cosine_decay", "mfu": 58.050105, "stage_name": "permutation_cycle", "status": "training", "step": 787, "tokens_per_sec_per_gpu": 101055.379028, "vram_gb": 20.709735} {"adam_lr": 7.590135046865651e-05, "dt_ms": 164.015293, "global_tokens_per_sec": 99893.123917, "grad_norm": 24.625, "loss": 3.055779, "lr_stage": "cosine_decay", "mfu": 57.38246, "stage_name": "permutation_cycle", "status": "training", "step": 788, "tokens_per_sec_per_gpu": 99893.123917, "vram_gb": 20.709735} {"adam_lr": 7.58266664433321e-05, "dt_ms": 163.686037, "global_tokens_per_sec": 100094.059908, "grad_norm": 25.375, "loss": 3.269886, "lr_stage": "cosine_decay", "mfu": 57.497886, "stage_name": "permutation_cycle", "status": "training", "step": 789, "tokens_per_sec_per_gpu": 100094.059908, "vram_gb": 20.709735} {"adam_lr": 7.575190374550272e-05, "dt_ms": 163.511038, "global_tokens_per_sec": 100201.186524, "grad_norm": 33.5, "loss": 3.016948, "lr_stage": "cosine_decay", "mfu": 57.559424, "stage_name": "permutation_cycle", "status": "training", "step": 790, "tokens_per_sec_per_gpu": 100201.186524, "vram_gb": 20.709735} {"adam_lr": 7.567706260290851e-05, "dt_ms": 163.495064, "global_tokens_per_sec": 100210.976534, "grad_norm": 37.0, "loss": 3.280713, "lr_stage": "cosine_decay", "mfu": 57.565047, "stage_name": "permutation_cycle", "status": "training", "step": 791, "tokens_per_sec_per_gpu": 100210.976534, "vram_gb": 20.709735} {"adam_lr": 7.560214324352858e-05, "dt_ms": 162.582874, "global_tokens_per_sec": 100773.221477, "grad_norm": 22.875, "loss": 3.269877, "lr_stage": "cosine_decay", "mfu": 57.888022, "stage_name": "permutation_cycle", "status": "training", "step": 792, "tokens_per_sec_per_gpu": 100773.221477, "vram_gb": 20.709735} {"adam_lr": 7.55271458955803e-05, "dt_ms": 163.190365, "global_tokens_per_sec": 100398.084264, "grad_norm": 29.5, "loss": 3.52428, "lr_stage": "cosine_decay", "mfu": 57.672529, "stage_name": "permutation_cycle", "status": "training", "step": 793, "tokens_per_sec_per_gpu": 100398.084264, "vram_gb": 20.709735} {"adam_lr": 7.545207078751857e-05, "dt_ms": 164.296389, "global_tokens_per_sec": 99722.216276, "grad_norm": 36.25, "loss": 3.232969, "lr_stage": "cosine_decay", "mfu": 57.284285, "stage_name": "permutation_cycle", "status": "training", "step": 794, "tokens_per_sec_per_gpu": 99722.216276, "vram_gb": 20.709735} {"adam_lr": 7.537691814803521e-05, "dt_ms": 161.635399, "global_tokens_per_sec": 101363.934603, "grad_norm": 21.75, "loss": 2.774686, "lr_stage": "cosine_decay", "mfu": 58.227351, "stage_name": "permutation_cycle", "status": "training", "step": 795, "tokens_per_sec_per_gpu": 101363.934603, "vram_gb": 20.709735} {"adam_lr": 7.530168820605818e-05, "dt_ms": 161.445141, "global_tokens_per_sec": 101483.388815, "grad_norm": 31.125, "loss": 2.962722, "lr_stage": "cosine_decay", "mfu": 58.29597, "stage_name": "permutation_cycle", "status": "training", "step": 796, "tokens_per_sec_per_gpu": 101483.388815, "vram_gb": 20.709735} {"adam_lr": 7.522638119075096e-05, "dt_ms": 164.556026, "global_tokens_per_sec": 99564.873755, "grad_norm": 31.125, "loss": 3.250639, "lr_stage": "cosine_decay", "mfu": 57.193901, "stage_name": "permutation_cycle", "status": "training", "step": 797, "tokens_per_sec_per_gpu": 99564.873755, "vram_gb": 20.709735} {"adam_lr": 7.515099733151177e-05, "dt_ms": 163.31768, "global_tokens_per_sec": 100319.818185, "grad_norm": 29.5, "loss": 3.044911, "lr_stage": "cosine_decay", "mfu": 57.62757, "stage_name": "permutation_cycle", "status": "training", "step": 798, "tokens_per_sec_per_gpu": 100319.818185, "vram_gb": 20.709735} {"adam_lr": 7.507553685797287e-05, "dt_ms": 161.516905, "global_tokens_per_sec": 101438.298469, "grad_norm": 32.5, "loss": 2.996252, "lr_stage": "cosine_decay", "mfu": 58.270068, "stage_name": "permutation_cycle", "status": "training", "step": 799, "tokens_per_sec_per_gpu": 101438.298469, "vram_gb": 20.709735} {"adam_lr": 7.500000000000001e-05, "dt_ms": 165.136576, "global_tokens_per_sec": 99214.846442, "grad_norm": 23.875, "loss": 2.570614, "lr_stage": "cosine_decay", "mfu": 56.992832, "stage_name": "permutation_cycle", "status": "training", "step": 800, "tokens_per_sec_per_gpu": 99214.846442, "vram_gb": 20.709735} {"adam_lr": 7.492438698769152e-05, "dt_ms": 164.189577, "global_tokens_per_sec": 99787.089346, "grad_norm": 40.75, "loss": 3.011016, "lr_stage": "cosine_decay", "mfu": 57.32155, "stage_name": "permutation_cycle", "status": "training", "step": 801, "tokens_per_sec_per_gpu": 99787.089346, "vram_gb": 20.709735} {"adam_lr": 7.484869805137778e-05, "dt_ms": 162.168264, "global_tokens_per_sec": 101030.864835, "grad_norm": 30.25, "loss": 3.169347, "lr_stage": "cosine_decay", "mfu": 58.036023, "stage_name": "permutation_cycle", "status": "training", "step": 802, "tokens_per_sec_per_gpu": 101030.864835, "vram_gb": 20.709735} {"adam_lr": 7.477293342162039e-05, "dt_ms": 163.787127, "global_tokens_per_sec": 100032.281816, "grad_norm": 33.5, "loss": 2.879724, "lr_stage": "cosine_decay", "mfu": 57.462398, "stage_name": "permutation_cycle", "status": "training", "step": 803, "tokens_per_sec_per_gpu": 100032.281816, "vram_gb": 20.709735} {"adam_lr": 7.469709332921155e-05, "dt_ms": 163.179636, "global_tokens_per_sec": 100404.685299, "grad_norm": 19.25, "loss": 2.898166, "lr_stage": "cosine_decay", "mfu": 57.676321, "stage_name": "permutation_cycle", "status": "training", "step": 804, "tokens_per_sec_per_gpu": 100404.685299, "vram_gb": 20.709735} {"adam_lr": 7.462117800517336e-05, "dt_ms": 162.922382, "global_tokens_per_sec": 100563.223808, "grad_norm": 47.0, "loss": 4.014221, "lr_stage": "cosine_decay", "mfu": 57.767392, "stage_name": "permutation_cycle", "status": "training", "step": 805, "tokens_per_sec_per_gpu": 100563.223808, "vram_gb": 20.709735} {"adam_lr": 7.454518768075704e-05, "dt_ms": 163.725615, "global_tokens_per_sec": 100069.864115, "grad_norm": 37.5, "loss": 3.568097, "lr_stage": "cosine_decay", "mfu": 57.483987, "stage_name": "permutation_cycle", "status": "training", "step": 806, "tokens_per_sec_per_gpu": 100069.864115, "vram_gb": 20.709735} {"adam_lr": 7.446912258744232e-05, "dt_ms": 162.849903, "global_tokens_per_sec": 100607.98126, "grad_norm": 29.0, "loss": 2.212195, "lr_stage": "cosine_decay", "mfu": 57.793102, "stage_name": "permutation_cycle", "status": "training", "step": 807, "tokens_per_sec_per_gpu": 100607.98126, "vram_gb": 20.709735} {"adam_lr": 7.439298295693665e-05, "dt_ms": 162.843466, "global_tokens_per_sec": 100611.958355, "grad_norm": 28.0, "loss": 3.318239, "lr_stage": "cosine_decay", "mfu": 57.795387, "stage_name": "permutation_cycle", "status": "training", "step": 808, "tokens_per_sec_per_gpu": 100611.958355, "vram_gb": 20.709735} {"adam_lr": 7.431676902117452e-05, "dt_ms": 164.089203, "global_tokens_per_sec": 99848.129629, "grad_norm": 34.25, "loss": 3.57376, "lr_stage": "cosine_decay", "mfu": 57.356614, "stage_name": "permutation_cycle", "status": "training", "step": 809, "tokens_per_sec_per_gpu": 99848.129629, "vram_gb": 20.709735} {"adam_lr": 7.424048101231686e-05, "dt_ms": 162.269831, "global_tokens_per_sec": 100967.628603, "grad_norm": 30.375, "loss": 2.837499, "lr_stage": "cosine_decay", "mfu": 57.999697, "stage_name": "permutation_cycle", "status": "training", "step": 810, "tokens_per_sec_per_gpu": 100967.628603, "vram_gb": 20.709735} {"adam_lr": 7.416411916275012e-05, "dt_ms": 163.614273, "global_tokens_per_sec": 100137.962859, "grad_norm": 54.5, "loss": 3.495548, "lr_stage": "cosine_decay", "mfu": 57.523105, "stage_name": "permutation_cycle", "status": "training", "step": 811, "tokens_per_sec_per_gpu": 100137.962859, "vram_gb": 20.709735} {"adam_lr": 7.408768370508576e-05, "dt_ms": 163.269043, "global_tokens_per_sec": 100349.703178, "grad_norm": 23.5, "loss": 3.206644, "lr_stage": "cosine_decay", "mfu": 57.644737, "stage_name": "permutation_cycle", "status": "training", "step": 812, "tokens_per_sec_per_gpu": 100349.703178, "vram_gb": 20.709735} {"adam_lr": 7.401117487215944e-05, "dt_ms": 163.622379, "global_tokens_per_sec": 100133.001792, "grad_norm": 23.75, "loss": 3.500627, "lr_stage": "cosine_decay", "mfu": 57.520256, "stage_name": "permutation_cycle", "status": "training", "step": 813, "tokens_per_sec_per_gpu": 100133.001792, "vram_gb": 20.709735} {"adam_lr": 7.393459289703035e-05, "dt_ms": 164.700985, "global_tokens_per_sec": 99477.243591, "grad_norm": 33.5, "loss": 3.209254, "lr_stage": "cosine_decay", "mfu": 57.143563, "stage_name": "permutation_cycle", "status": "training", "step": 814, "tokens_per_sec_per_gpu": 99477.243591, "vram_gb": 20.709735} {"adam_lr": 7.385793801298042e-05, "dt_ms": 162.985086, "global_tokens_per_sec": 100524.534838, "grad_norm": 32.5, "loss": 3.90705, "lr_stage": "cosine_decay", "mfu": 57.745167, "stage_name": "permutation_cycle", "status": "training", "step": 815, "tokens_per_sec_per_gpu": 100524.534838, "vram_gb": 20.709735} {"adam_lr": 7.378121045351378e-05, "dt_ms": 163.828373, "global_tokens_per_sec": 100007.097089, "grad_norm": 40.5, "loss": 3.212194, "lr_stage": "cosine_decay", "mfu": 57.447931, "stage_name": "permutation_cycle", "status": "training", "step": 816, "tokens_per_sec_per_gpu": 100007.097089, "vram_gb": 20.709735} {"adam_lr": 7.37044104523558e-05, "dt_ms": 163.031101, "global_tokens_per_sec": 100496.162246, "grad_norm": 21.5, "loss": 2.946869, "lr_stage": "cosine_decay", "mfu": 57.728869, "stage_name": "permutation_cycle", "status": "training", "step": 817, "tokens_per_sec_per_gpu": 100496.162246, "vram_gb": 20.709735} {"adam_lr": 7.362753824345272e-05, "dt_ms": 163.308382, "global_tokens_per_sec": 100325.530116, "grad_norm": 48.75, "loss": 3.880997, "lr_stage": "cosine_decay", "mfu": 57.630851, "stage_name": "permutation_cycle", "status": "training", "step": 818, "tokens_per_sec_per_gpu": 100325.530116, "vram_gb": 20.709735} {"adam_lr": 7.35505940609705e-05, "dt_ms": 162.558794, "global_tokens_per_sec": 100788.149288, "grad_norm": 41.5, "loss": 3.881939, "lr_stage": "cosine_decay", "mfu": 57.896598, "stage_name": "permutation_cycle", "status": "training", "step": 819, "tokens_per_sec_per_gpu": 100788.149288, "vram_gb": 20.709735} {"adam_lr": 7.347357813929454e-05, "dt_ms": 164.680481, "global_tokens_per_sec": 99489.62928, "grad_norm": 49.25, "loss": 3.008103, "lr_stage": "cosine_decay", "mfu": 57.150678, "stage_name": "permutation_cycle", "status": "training", "step": 820, "tokens_per_sec_per_gpu": 99489.62928, "vram_gb": 20.709735} {"adam_lr": 7.339649071302867e-05, "dt_ms": 161.854506, "global_tokens_per_sec": 101226.715595, "grad_norm": 54.25, "loss": 2.793291, "lr_stage": "cosine_decay", "mfu": 58.148527, "stage_name": "permutation_cycle", "status": "training", "step": 821, "tokens_per_sec_per_gpu": 101226.715595, "vram_gb": 20.709735} {"adam_lr": 7.331933201699457e-05, "dt_ms": 163.26046, "global_tokens_per_sec": 100354.978848, "grad_norm": 23.0, "loss": 2.909181, "lr_stage": "cosine_decay", "mfu": 57.647768, "stage_name": "permutation_cycle", "status": "training", "step": 822, "tokens_per_sec_per_gpu": 100354.978848, "vram_gb": 20.709735} {"adam_lr": 7.3242102286231e-05, "dt_ms": 163.787603, "global_tokens_per_sec": 100031.990591, "grad_norm": 23.0, "loss": 2.503188, "lr_stage": "cosine_decay", "mfu": 57.462231, "stage_name": "permutation_cycle", "status": "training", "step": 823, "tokens_per_sec_per_gpu": 100031.990591, "vram_gb": 20.709735} {"adam_lr": 7.316480175599309e-05, "dt_ms": 162.365437, "global_tokens_per_sec": 100908.175704, "grad_norm": 46.5, "loss": 3.161757, "lr_stage": "cosine_decay", "mfu": 57.965545, "stage_name": "permutation_cycle", "status": "training", "step": 824, "tokens_per_sec_per_gpu": 100908.175704, "vram_gb": 20.709735} {"adam_lr": 7.308743066175172e-05, "dt_ms": 162.93478, "global_tokens_per_sec": 100555.571916, "grad_norm": 31.875, "loss": 2.283397, "lr_stage": "cosine_decay", "mfu": 57.762996, "stage_name": "permutation_cycle", "status": "training", "step": 825, "tokens_per_sec_per_gpu": 100555.571916, "vram_gb": 20.709735} {"adam_lr": 7.300998923919259e-05, "dt_ms": 164.755106, "global_tokens_per_sec": 99444.565941, "grad_norm": 29.25, "loss": 2.721467, "lr_stage": "cosine_decay", "mfu": 57.124791, "stage_name": "permutation_cycle", "status": "training", "step": 826, "tokens_per_sec_per_gpu": 99444.565941, "vram_gb": 20.709735} {"adam_lr": 7.293247772421576e-05, "dt_ms": 162.235975, "global_tokens_per_sec": 100988.69855, "grad_norm": 42.25, "loss": 2.742181, "lr_stage": "cosine_decay", "mfu": 58.011801, "stage_name": "permutation_cycle", "status": "training", "step": 827, "tokens_per_sec_per_gpu": 100988.69855, "vram_gb": 20.709735} {"adam_lr": 7.285489635293472e-05, "dt_ms": 162.318707, "global_tokens_per_sec": 100937.226226, "grad_norm": 36.25, "loss": 3.182179, "lr_stage": "cosine_decay", "mfu": 57.982233, "stage_name": "permutation_cycle", "status": "training", "step": 828, "tokens_per_sec_per_gpu": 100937.226226, "vram_gb": 20.709735} {"adam_lr": 7.277724536167579e-05, "dt_ms": 163.454294, "global_tokens_per_sec": 100235.97165, "grad_norm": 26.625, "loss": 3.452477, "lr_stage": "cosine_decay", "mfu": 57.579405, "stage_name": "permutation_cycle", "status": "training", "step": 829, "tokens_per_sec_per_gpu": 100235.97165, "vram_gb": 20.709735} {"adam_lr": 7.269952498697734e-05, "dt_ms": 162.035942, "global_tokens_per_sec": 101113.368984, "grad_norm": 29.0, "loss": 2.372988, "lr_stage": "cosine_decay", "mfu": 58.083416, "stage_name": "permutation_cycle", "status": "training", "step": 830, "tokens_per_sec_per_gpu": 101113.368984, "vram_gb": 20.709735} {"adam_lr": 7.262173546558914e-05, "dt_ms": 163.754463, "global_tokens_per_sec": 100052.234793, "grad_norm": 36.0, "loss": 2.62887, "lr_stage": "cosine_decay", "mfu": 57.47386, "stage_name": "permutation_cycle", "status": "training", "step": 831, "tokens_per_sec_per_gpu": 100052.234793, "vram_gb": 20.709735} {"adam_lr": 7.254387703447154e-05, "dt_ms": 163.904667, "global_tokens_per_sec": 99960.546028, "grad_norm": 29.375, "loss": 2.925132, "lr_stage": "cosine_decay", "mfu": 57.42119, "stage_name": "permutation_cycle", "status": "training", "step": 832, "tokens_per_sec_per_gpu": 99960.546028, "vram_gb": 20.709735} {"adam_lr": 7.246594993079482e-05, "dt_ms": 163.721561, "global_tokens_per_sec": 100072.34146, "grad_norm": 32.75, "loss": 2.528666, "lr_stage": "cosine_decay", "mfu": 57.48541, "stage_name": "permutation_cycle", "status": "training", "step": 833, "tokens_per_sec_per_gpu": 100072.34146, "vram_gb": 20.709735} {"adam_lr": 7.238795439193848e-05, "dt_ms": 163.645744, "global_tokens_per_sec": 100118.704997, "grad_norm": 36.25, "loss": 2.83648, "lr_stage": "cosine_decay", "mfu": 57.512043, "stage_name": "permutation_cycle", "status": "training", "step": 834, "tokens_per_sec_per_gpu": 100118.704997, "vram_gb": 20.709735} {"adam_lr": 7.230989065549044e-05, "dt_ms": 164.064407, "global_tokens_per_sec": 99863.219968, "grad_norm": 23.0, "loss": 2.661124, "lr_stage": "cosine_decay", "mfu": 57.365283, "stage_name": "permutation_cycle", "status": "training", "step": 835, "tokens_per_sec_per_gpu": 99863.219968, "vram_gb": 20.709735} {"adam_lr": 7.223175895924638e-05, "dt_ms": 163.660049, "global_tokens_per_sec": 100109.953872, "grad_norm": 39.0, "loss": 3.728054, "lr_stage": "cosine_decay", "mfu": 57.507016, "stage_name": "permutation_cycle", "status": "training", "step": 836, "tokens_per_sec_per_gpu": 100109.953872, "vram_gb": 20.709735} {"adam_lr": 7.215355954120899e-05, "dt_ms": 163.049459, "global_tokens_per_sec": 100484.84708, "grad_norm": 39.25, "loss": 2.964264, "lr_stage": "cosine_decay", "mfu": 57.722369, "stage_name": "permutation_cycle", "status": "training", "step": 837, "tokens_per_sec_per_gpu": 100484.84708, "vram_gb": 20.709735} {"adam_lr": 7.207529263958726e-05, "dt_ms": 164.111137, "global_tokens_per_sec": 99834.784284, "grad_norm": 38.25, "loss": 2.777285, "lr_stage": "cosine_decay", "mfu": 57.348948, "stage_name": "permutation_cycle", "status": "training", "step": 838, "tokens_per_sec_per_gpu": 99834.784284, "vram_gb": 20.709735} {"adam_lr": 7.199695849279576e-05, "dt_ms": 163.244724, "global_tokens_per_sec": 100364.652352, "grad_norm": 28.0, "loss": 2.546182, "lr_stage": "cosine_decay", "mfu": 57.653325, "stage_name": "permutation_cycle", "status": "training", "step": 839, "tokens_per_sec_per_gpu": 100364.652352, "vram_gb": 20.709735} {"adam_lr": 7.191855733945387e-05, "dt_ms": 163.667917, "global_tokens_per_sec": 100105.141405, "grad_norm": 29.125, "loss": 3.055521, "lr_stage": "cosine_decay", "mfu": 57.504252, "stage_name": "permutation_cycle", "status": "training", "step": 840, "tokens_per_sec_per_gpu": 100105.141405, "vram_gb": 20.709735} {"adam_lr": 7.184008941838512e-05, "dt_ms": 164.115906, "global_tokens_per_sec": 99831.883594, "grad_norm": 42.25, "loss": 3.642052, "lr_stage": "cosine_decay", "mfu": 57.347282, "stage_name": "permutation_cycle", "status": "training", "step": 841, "tokens_per_sec_per_gpu": 99831.883594, "vram_gb": 20.709735} {"adam_lr": 7.176155496861638e-05, "dt_ms": 162.106276, "global_tokens_per_sec": 101069.498658, "grad_norm": 40.5, "loss": 2.340189, "lr_stage": "cosine_decay", "mfu": 58.058215, "stage_name": "permutation_cycle", "status": "training", "step": 842, "tokens_per_sec_per_gpu": 101069.498658, "vram_gb": 20.709735} {"adam_lr": 7.168295422937723e-05, "dt_ms": 164.053679, "global_tokens_per_sec": 99869.750856, "grad_norm": 41.0, "loss": 3.294841, "lr_stage": "cosine_decay", "mfu": 57.369034, "stage_name": "permutation_cycle", "status": "training", "step": 843, "tokens_per_sec_per_gpu": 99869.750856, "vram_gb": 20.709735} {"adam_lr": 7.160428744009912e-05, "dt_ms": 164.699793, "global_tokens_per_sec": 99477.963605, "grad_norm": 38.75, "loss": 3.670086, "lr_stage": "cosine_decay", "mfu": 57.143976, "stage_name": "permutation_cycle", "status": "training", "step": 844, "tokens_per_sec_per_gpu": 99477.963605, "vram_gb": 20.709735} {"adam_lr": 7.152555484041476e-05, "dt_ms": 163.338184, "global_tokens_per_sec": 100307.224943, "grad_norm": 26.375, "loss": 3.566465, "lr_stage": "cosine_decay", "mfu": 57.620336, "stage_name": "permutation_cycle", "status": "training", "step": 845, "tokens_per_sec_per_gpu": 100307.224943, "vram_gb": 20.709735} {"adam_lr": 7.14467566701573e-05, "dt_ms": 161.526918, "global_tokens_per_sec": 101432.009978, "grad_norm": 43.0, "loss": 1.730645, "lr_stage": "cosine_decay", "mfu": 58.266456, "stage_name": "permutation_cycle", "status": "training", "step": 846, "tokens_per_sec_per_gpu": 101432.009978, "vram_gb": 20.709735} {"adam_lr": 7.136789316935964e-05, "dt_ms": 163.878918, "global_tokens_per_sec": 99976.252166, "grad_norm": 27.25, "loss": 2.837263, "lr_stage": "cosine_decay", "mfu": 57.430213, "stage_name": "permutation_cycle", "status": "training", "step": 847, "tokens_per_sec_per_gpu": 99976.252166, "vram_gb": 20.709735} {"adam_lr": 7.128896457825364e-05, "dt_ms": 161.789656, "global_tokens_per_sec": 101267.290116, "grad_norm": 34.75, "loss": 3.361201, "lr_stage": "cosine_decay", "mfu": 58.171835, "stage_name": "permutation_cycle", "status": "training", "step": 848, "tokens_per_sec_per_gpu": 101267.290116, "vram_gb": 20.709735} {"adam_lr": 7.120997113726951e-05, "dt_ms": 162.556171, "global_tokens_per_sec": 100789.775357, "grad_norm": 28.125, "loss": 3.122241, "lr_stage": "cosine_decay", "mfu": 57.897532, "stage_name": "permutation_cycle", "status": "training", "step": 849, "tokens_per_sec_per_gpu": 100789.775357, "vram_gb": 20.709735} {"adam_lr": 7.113091308703498e-05, "dt_ms": 162.542582, "global_tokens_per_sec": 100798.202188, "grad_norm": 30.625, "loss": 3.455907, "lr_stage": "cosine_decay", "mfu": 57.902372, "stage_name": "permutation_cycle", "status": "training", "step": 850, "tokens_per_sec_per_gpu": 100798.202188, "vram_gb": 20.709735} {"adam_lr": 7.105179066837456e-05, "dt_ms": 164.009571, "global_tokens_per_sec": 99896.609037, "grad_norm": 36.5, "loss": 3.503568, "lr_stage": "cosine_decay", "mfu": 57.384462, "stage_name": "permutation_cycle", "status": "training", "step": 851, "tokens_per_sec_per_gpu": 99896.609037, "vram_gb": 20.709735} {"adam_lr": 7.097260412230886e-05, "dt_ms": 162.985086, "global_tokens_per_sec": 100524.534838, "grad_norm": 33.0, "loss": 3.391824, "lr_stage": "cosine_decay", "mfu": 57.745167, "stage_name": "permutation_cycle", "status": "training", "step": 852, "tokens_per_sec_per_gpu": 100524.534838, "vram_gb": 20.709735} {"adam_lr": 7.089335369005385e-05, "dt_ms": 163.744926, "global_tokens_per_sec": 100058.061981, "grad_norm": 38.75, "loss": 2.855662, "lr_stage": "cosine_decay", "mfu": 57.477207, "stage_name": "permutation_cycle", "status": "training", "step": 853, "tokens_per_sec_per_gpu": 100058.061981, "vram_gb": 20.709735} {"adam_lr": 7.081403961302006e-05, "dt_ms": 161.626577, "global_tokens_per_sec": 101369.466989, "grad_norm": 45.0, "loss": 3.132751, "lr_stage": "cosine_decay", "mfu": 58.230529, "stage_name": "permutation_cycle", "status": "training", "step": 854, "tokens_per_sec_per_gpu": 101369.466989, "vram_gb": 20.709735} {"adam_lr": 7.073466213281196e-05, "dt_ms": 165.809631, "global_tokens_per_sec": 98812.112824, "grad_norm": 30.875, "loss": 3.623363, "lr_stage": "cosine_decay", "mfu": 56.761486, "stage_name": "permutation_cycle", "status": "training", "step": 855, "tokens_per_sec_per_gpu": 98812.112824, "vram_gb": 20.709735} {"adam_lr": 7.06552214912271e-05, "dt_ms": 162.832975, "global_tokens_per_sec": 100618.440221, "grad_norm": 31.375, "loss": 3.191919, "lr_stage": "cosine_decay", "mfu": 57.79911, "stage_name": "permutation_cycle", "status": "training", "step": 856, "tokens_per_sec_per_gpu": 100618.440221, "vram_gb": 20.709735} {"adam_lr": 7.057571793025544e-05, "dt_ms": 164.244175, "global_tokens_per_sec": 99753.918239, "grad_norm": 22.625, "loss": 3.029161, "lr_stage": "cosine_decay", "mfu": 57.302495, "stage_name": "permutation_cycle", "status": "training", "step": 857, "tokens_per_sec_per_gpu": 99753.918239, "vram_gb": 20.709735} {"adam_lr": 7.049615169207864e-05, "dt_ms": 163.596392, "global_tokens_per_sec": 100148.908127, "grad_norm": 41.75, "loss": 3.488114, "lr_stage": "cosine_decay", "mfu": 57.529393, "stage_name": "permutation_cycle", "status": "training", "step": 858, "tokens_per_sec_per_gpu": 100148.908127, "vram_gb": 20.709735} {"adam_lr": 7.041652301906924e-05, "dt_ms": 161.497593, "global_tokens_per_sec": 101450.428475, "grad_norm": 24.75, "loss": 2.20485, "lr_stage": "cosine_decay", "mfu": 58.277036, "stage_name": "permutation_cycle", "status": "training", "step": 859, "tokens_per_sec_per_gpu": 101450.428475, "vram_gb": 20.709735} {"adam_lr": 7.033683215379002e-05, "dt_ms": 160.514593, "global_tokens_per_sec": 102071.716229, "grad_norm": 24.25, "loss": 2.624678, "lr_stage": "cosine_decay", "mfu": 58.633928, "stage_name": "permutation_cycle", "status": "training", "step": 860, "tokens_per_sec_per_gpu": 102071.716229, "vram_gb": 20.709735} {"adam_lr": 7.025707933899314e-05, "dt_ms": 163.340092, "global_tokens_per_sec": 100306.053639, "grad_norm": 30.875, "loss": 3.296671, "lr_stage": "cosine_decay", "mfu": 57.619663, "stage_name": "permutation_cycle", "status": "training", "step": 861, "tokens_per_sec_per_gpu": 100306.053639, "vram_gb": 20.709735} {"adam_lr": 7.017726481761951e-05, "dt_ms": 165.20381, "global_tokens_per_sec": 99174.468349, "grad_norm": 34.0, "loss": 2.345993, "lr_stage": "cosine_decay", "mfu": 56.969637, "stage_name": "permutation_cycle", "status": "training", "step": 862, "tokens_per_sec_per_gpu": 99174.468349, "vram_gb": 20.709735} {"adam_lr": 7.009738883279802e-05, "dt_ms": 161.846399, "global_tokens_per_sec": 101231.785632, "grad_norm": 54.5, "loss": 3.223689, "lr_stage": "cosine_decay", "mfu": 58.151439, "stage_name": "permutation_cycle", "status": "training", "step": 863, "tokens_per_sec_per_gpu": 101231.785632, "vram_gb": 20.709735} {"adam_lr": 7.001745162784477e-05, "dt_ms": 162.387848, "global_tokens_per_sec": 100894.249242, "grad_norm": 35.5, "loss": 2.505919, "lr_stage": "cosine_decay", "mfu": 57.957545, "stage_name": "permutation_cycle", "status": "training", "step": 864, "tokens_per_sec_per_gpu": 100894.249242, "vram_gb": 20.709735} {"adam_lr": 6.993745344626231e-05, "dt_ms": 162.724733, "global_tokens_per_sec": 100685.370087, "grad_norm": 37.25, "loss": 2.969539, "lr_stage": "cosine_decay", "mfu": 57.837557, "stage_name": "permutation_cycle", "status": "training", "step": 865, "tokens_per_sec_per_gpu": 100685.370087, "vram_gb": 20.709735} {"adam_lr": 6.985739453173903e-05, "dt_ms": 159.788132, "global_tokens_per_sec": 102535.775494, "grad_norm": 36.0, "loss": 2.865234, "lr_stage": "cosine_decay", "mfu": 58.900501, "stage_name": "permutation_cycle", "status": "training", "step": 866, "tokens_per_sec_per_gpu": 102535.775494, "vram_gb": 20.709735} {"adam_lr": 6.977727512814826e-05, "dt_ms": 163.777113, "global_tokens_per_sec": 100038.397941, "grad_norm": 33.0, "loss": 3.096727, "lr_stage": "cosine_decay", "mfu": 57.465911, "stage_name": "permutation_cycle", "status": "training", "step": 867, "tokens_per_sec_per_gpu": 100038.397941, "vram_gb": 20.709735} {"adam_lr": 6.969709547954756e-05, "dt_ms": 163.280487, "global_tokens_per_sec": 100342.669813, "grad_norm": 25.5, "loss": 3.022876, "lr_stage": "cosine_decay", "mfu": 57.640697, "stage_name": "permutation_cycle", "status": "training", "step": 868, "tokens_per_sec_per_gpu": 100342.669813, "vram_gb": 20.709735} {"adam_lr": 6.961685583017808e-05, "dt_ms": 162.085295, "global_tokens_per_sec": 101082.581415, "grad_norm": 27.125, "loss": 2.849036, "lr_stage": "cosine_decay", "mfu": 58.065731, "stage_name": "permutation_cycle", "status": "training", "step": 869, "tokens_per_sec_per_gpu": 101082.581415, "vram_gb": 20.709735} {"adam_lr": 6.953655642446368e-05, "dt_ms": 162.639618, "global_tokens_per_sec": 100738.062531, "grad_norm": 36.0, "loss": 2.875651, "lr_stage": "cosine_decay", "mfu": 57.867826, "stage_name": "permutation_cycle", "status": "training", "step": 870, "tokens_per_sec_per_gpu": 100738.062531, "vram_gb": 20.709735} {"adam_lr": 6.94561975070103e-05, "dt_ms": 163.918734, "global_tokens_per_sec": 99951.967908, "grad_norm": 30.375, "loss": 3.255753, "lr_stage": "cosine_decay", "mfu": 57.416263, "stage_name": "permutation_cycle", "status": "training", "step": 871, "tokens_per_sec_per_gpu": 99951.967908, "vram_gb": 20.709735} {"adam_lr": 6.937577932260515e-05, "dt_ms": 161.092758, "global_tokens_per_sec": 101705.378847, "grad_norm": 32.5, "loss": 2.432382, "lr_stage": "cosine_decay", "mfu": 58.42349, "stage_name": "permutation_cycle", "status": "training", "step": 872, "tokens_per_sec_per_gpu": 101705.378847, "vram_gb": 20.709735} {"adam_lr": 6.929530211621594e-05, "dt_ms": 164.093018, "global_tokens_per_sec": 99845.808443, "grad_norm": 25.875, "loss": 2.784633, "lr_stage": "cosine_decay", "mfu": 57.355281, "stage_name": "permutation_cycle", "status": "training", "step": 873, "tokens_per_sec_per_gpu": 99845.808443, "vram_gb": 20.709735} {"adam_lr": 6.921476613299018e-05, "dt_ms": 162.545443, "global_tokens_per_sec": 100796.428001, "grad_norm": 34.75, "loss": 2.974927, "lr_stage": "cosine_decay", "mfu": 57.901353, "stage_name": "permutation_cycle", "status": "training", "step": 874, "tokens_per_sec_per_gpu": 100796.428001, "vram_gb": 20.709735} {"adam_lr": 6.91341716182545e-05, "dt_ms": 162.822723, "global_tokens_per_sec": 100624.775578, "grad_norm": 29.375, "loss": 3.242425, "lr_stage": "cosine_decay", "mfu": 57.802749, "stage_name": "permutation_cycle", "status": "training", "step": 875, "tokens_per_sec_per_gpu": 100624.775578, "vram_gb": 20.709735} {"adam_lr": 6.905351881751372e-05, "dt_ms": 164.151669, "global_tokens_per_sec": 99810.133792, "grad_norm": 36.75, "loss": 3.329122, "lr_stage": "cosine_decay", "mfu": 57.334788, "stage_name": "permutation_cycle", "status": "training", "step": 876, "tokens_per_sec_per_gpu": 99810.133792, "vram_gb": 20.709735} {"adam_lr": 6.897280797645026e-05, "dt_ms": 163.827181, "global_tokens_per_sec": 100007.824793, "grad_norm": 41.5, "loss": 3.666473, "lr_stage": "cosine_decay", "mfu": 57.448349, "stage_name": "permutation_cycle", "status": "training", "step": 877, "tokens_per_sec_per_gpu": 100007.824793, "vram_gb": 20.709735} {"adam_lr": 6.889203934092336e-05, "dt_ms": 161.435127, "global_tokens_per_sec": 101489.683678, "grad_norm": 35.0, "loss": 3.564422, "lr_stage": "cosine_decay", "mfu": 58.299586, "stage_name": "permutation_cycle", "status": "training", "step": 878, "tokens_per_sec_per_gpu": 101489.683678, "vram_gb": 20.709735} {"adam_lr": 6.881121315696827e-05, "dt_ms": 161.800146, "global_tokens_per_sec": 101260.724385, "grad_norm": 31.625, "loss": 3.183656, "lr_stage": "cosine_decay", "mfu": 58.168063, "stage_name": "permutation_cycle", "status": "training", "step": 879, "tokens_per_sec_per_gpu": 101260.724385, "vram_gb": 20.709735} {"adam_lr": 6.873032967079561e-05, "dt_ms": 162.278652, "global_tokens_per_sec": 100962.139991, "grad_norm": 26.375, "loss": 2.782424, "lr_stage": "cosine_decay", "mfu": 57.996545, "stage_name": "permutation_cycle", "status": "training", "step": 880, "tokens_per_sec_per_gpu": 100962.139991, "vram_gb": 20.709735} {"adam_lr": 6.864938912879046e-05, "dt_ms": 163.346767, "global_tokens_per_sec": 100301.954292, "grad_norm": 25.75, "loss": 2.084785, "lr_stage": "cosine_decay", "mfu": 57.617308, "stage_name": "permutation_cycle", "status": "training", "step": 881, "tokens_per_sec_per_gpu": 100301.954292, "vram_gb": 20.709735} {"adam_lr": 6.856839177751176e-05, "dt_ms": 163.778543, "global_tokens_per_sec": 100037.524163, "grad_norm": 26.625, "loss": 2.839528, "lr_stage": "cosine_decay", "mfu": 57.46541, "stage_name": "permutation_cycle", "status": "training", "step": 882, "tokens_per_sec_per_gpu": 100037.524163, "vram_gb": 20.709735} {"adam_lr": 6.848733786369148e-05, "dt_ms": 163.177013, "global_tokens_per_sec": 100406.299018, "grad_norm": 31.5, "loss": 2.510356, "lr_stage": "cosine_decay", "mfu": 57.677248, "stage_name": "permutation_cycle", "status": "training", "step": 883, "tokens_per_sec_per_gpu": 100406.299018, "vram_gb": 20.709735} {"adam_lr": 6.840622763423391e-05, "dt_ms": 163.005352, "global_tokens_per_sec": 100512.037163, "grad_norm": 28.75, "loss": 2.394905, "lr_stage": "cosine_decay", "mfu": 57.737988, "stage_name": "permutation_cycle", "status": "training", "step": 884, "tokens_per_sec_per_gpu": 100512.037163, "vram_gb": 20.709735} {"adam_lr": 6.832506133621487e-05, "dt_ms": 163.462162, "global_tokens_per_sec": 100231.14706, "grad_norm": 33.75, "loss": 3.275703, "lr_stage": "cosine_decay", "mfu": 57.576634, "stage_name": "permutation_cycle", "status": "training", "step": 885, "tokens_per_sec_per_gpu": 100231.14706, "vram_gb": 20.709735} {"adam_lr": 6.824383921688098e-05, "dt_ms": 162.994862, "global_tokens_per_sec": 100518.50616, "grad_norm": 33.75, "loss": 3.158723, "lr_stage": "cosine_decay", "mfu": 57.741704, "stage_name": "permutation_cycle", "status": "training", "step": 886, "tokens_per_sec_per_gpu": 100518.50616, "vram_gb": 20.709735} {"adam_lr": 6.816256152364892e-05, "dt_ms": 163.337469, "global_tokens_per_sec": 100307.664189, "grad_norm": 33.75, "loss": 4.017072, "lr_stage": "cosine_decay", "mfu": 57.620588, "stage_name": "permutation_cycle", "status": "training", "step": 887, "tokens_per_sec_per_gpu": 100307.664189, "vram_gb": 20.709735} {"adam_lr": 6.808122850410461e-05, "dt_ms": 164.613008, "global_tokens_per_sec": 99530.408619, "grad_norm": 24.375, "loss": 2.532896, "lr_stage": "cosine_decay", "mfu": 57.174103, "stage_name": "permutation_cycle", "status": "training", "step": 888, "tokens_per_sec_per_gpu": 99530.408619, "vram_gb": 20.709735} {"adam_lr": 6.799984040600257e-05, "dt_ms": 164.914846, "global_tokens_per_sec": 99348.241566, "grad_norm": 26.375, "loss": 2.98048, "lr_stage": "cosine_decay", "mfu": 57.069459, "stage_name": "permutation_cycle", "status": "training", "step": 889, "tokens_per_sec_per_gpu": 99348.241566, "vram_gb": 20.709735} {"adam_lr": 6.7918397477265e-05, "dt_ms": 162.832737, "global_tokens_per_sec": 100618.587546, "grad_norm": 39.5, "loss": 3.174558, "lr_stage": "cosine_decay", "mfu": 57.799195, "stage_name": "permutation_cycle", "status": "training", "step": 890, "tokens_per_sec_per_gpu": 100618.587546, "vram_gb": 20.709735} {"adam_lr": 6.783689996598126e-05, "dt_ms": 161.011934, "global_tokens_per_sec": 101756.432362, "grad_norm": 28.875, "loss": 2.172134, "lr_stage": "cosine_decay", "mfu": 58.452817, "stage_name": "permutation_cycle", "status": "training", "step": 891, "tokens_per_sec_per_gpu": 101756.432362, "vram_gb": 20.709735} {"adam_lr": 6.775534812040685e-05, "dt_ms": 163.605452, "global_tokens_per_sec": 100143.362225, "grad_norm": 37.25, "loss": 2.978508, "lr_stage": "cosine_decay", "mfu": 57.526207, "stage_name": "permutation_cycle", "status": "training", "step": 892, "tokens_per_sec_per_gpu": 100143.362225, "vram_gb": 20.709735} {"adam_lr": 6.767374218896286e-05, "dt_ms": 162.590265, "global_tokens_per_sec": 100768.64056, "grad_norm": 36.75, "loss": 2.966429, "lr_stage": "cosine_decay", "mfu": 57.885391, "stage_name": "permutation_cycle", "status": "training", "step": 893, "tokens_per_sec_per_gpu": 100768.64056, "vram_gb": 20.709735} {"adam_lr": 6.759208242023509e-05, "dt_ms": 164.38365, "global_tokens_per_sec": 99669.279867, "grad_norm": 23.875, "loss": 2.920849, "lr_stage": "cosine_decay", "mfu": 57.253876, "stage_name": "permutation_cycle", "status": "training", "step": 894, "tokens_per_sec_per_gpu": 99669.279867, "vram_gb": 20.709735} {"adam_lr": 6.751036906297337e-05, "dt_ms": 163.19418, "global_tokens_per_sec": 100395.737438, "grad_norm": 25.625, "loss": 2.774561, "lr_stage": "cosine_decay", "mfu": 57.671181, "stage_name": "permutation_cycle", "status": "training", "step": 895, "tokens_per_sec_per_gpu": 100395.737438, "vram_gb": 20.709735} {"adam_lr": 6.742860236609077e-05, "dt_ms": 163.464069, "global_tokens_per_sec": 100229.977533, "grad_norm": 30.375, "loss": 3.576801, "lr_stage": "cosine_decay", "mfu": 57.575962, "stage_name": "permutation_cycle", "status": "training", "step": 896, "tokens_per_sec_per_gpu": 100229.977533, "vram_gb": 20.709735} {"adam_lr": 6.734678257866279e-05, "dt_ms": 160.832882, "global_tokens_per_sec": 101869.715966, "grad_norm": 28.375, "loss": 1.829822, "lr_stage": "cosine_decay", "mfu": 58.517891, "stage_name": "permutation_cycle", "status": "training", "step": 897, "tokens_per_sec_per_gpu": 101869.715966, "vram_gb": 20.709735} {"adam_lr": 6.726490994992674e-05, "dt_ms": 163.682938, "global_tokens_per_sec": 100095.955254, "grad_norm": 29.0, "loss": 2.894229, "lr_stage": "cosine_decay", "mfu": 57.498975, "stage_name": "permutation_cycle", "status": "training", "step": 898, "tokens_per_sec_per_gpu": 100095.955254, "vram_gb": 20.709735} {"adam_lr": 6.718298472928082e-05, "dt_ms": 161.797523, "global_tokens_per_sec": 101262.365738, "grad_norm": 47.0, "loss": 3.147875, "lr_stage": "cosine_decay", "mfu": 58.169006, "stage_name": "permutation_cycle", "status": "training", "step": 899, "tokens_per_sec_per_gpu": 101262.365738, "vram_gb": 20.709735} {"adam_lr": 6.710100716628344e-05, "dt_ms": 162.65893, "global_tokens_per_sec": 100726.102266, "grad_norm": 37.5, "loss": 2.579081, "lr_stage": "cosine_decay", "mfu": 57.860955, "stage_name": "permutation_cycle", "status": "training", "step": 900, "tokens_per_sec_per_gpu": 100726.102266, "vram_gb": 20.709735} {"adam_lr": 6.701897751065252e-05, "dt_ms": 163.347006, "global_tokens_per_sec": 100301.807893, "grad_norm": 34.5, "loss": 3.028593, "lr_stage": "cosine_decay", "mfu": 57.617224, "stage_name": "permutation_cycle", "status": "training", "step": 901, "tokens_per_sec_per_gpu": 100301.807893, "vram_gb": 20.709735} {"adam_lr": 6.693689601226458e-05, "dt_ms": 163.507462, "global_tokens_per_sec": 100203.378151, "grad_norm": 41.5, "loss": 3.450403, "lr_stage": "cosine_decay", "mfu": 57.560682, "stage_name": "permutation_cycle", "status": "training", "step": 902, "tokens_per_sec_per_gpu": 100203.378151, "vram_gb": 20.709735} {"adam_lr": 6.685476292115412e-05, "dt_ms": 163.627148, "global_tokens_per_sec": 100130.083747, "grad_norm": 34.0, "loss": 4.075686, "lr_stage": "cosine_decay", "mfu": 57.518579, "stage_name": "permutation_cycle", "status": "training", "step": 903, "tokens_per_sec_per_gpu": 100130.083747, "vram_gb": 20.709735} {"adam_lr": 6.677257848751277e-05, "dt_ms": 162.465572, "global_tokens_per_sec": 100845.980858, "grad_norm": 32.5, "loss": 2.421301, "lr_stage": "cosine_decay", "mfu": 57.929818, "stage_name": "permutation_cycle", "status": "training", "step": 904, "tokens_per_sec_per_gpu": 100845.980858, "vram_gb": 20.709735} {"adam_lr": 6.669034296168855e-05, "dt_ms": 161.94725, "global_tokens_per_sec": 101168.744532, "grad_norm": 32.75, "loss": 3.489755, "lr_stage": "cosine_decay", "mfu": 58.115226, "stage_name": "permutation_cycle", "status": "training", "step": 905, "tokens_per_sec_per_gpu": 101168.744532, "vram_gb": 20.709735} {"adam_lr": 6.660805659418516e-05, "dt_ms": 165.022612, "global_tokens_per_sec": 99283.363894, "grad_norm": 29.625, "loss": 3.192012, "lr_stage": "cosine_decay", "mfu": 57.032191, "stage_name": "permutation_cycle", "status": "training", "step": 906, "tokens_per_sec_per_gpu": 99283.363894, "vram_gb": 20.709735} {"adam_lr": 6.652571963566115e-05, "dt_ms": 161.635637, "global_tokens_per_sec": 101363.785087, "grad_norm": 33.0, "loss": 2.3061, "lr_stage": "cosine_decay", "mfu": 58.227265, "stage_name": "permutation_cycle", "status": "training", "step": 907, "tokens_per_sec_per_gpu": 101363.785087, "vram_gb": 20.709735} {"adam_lr": 6.644333233692916e-05, "dt_ms": 163.090706, "global_tokens_per_sec": 100459.433985, "grad_norm": 35.25, "loss": 3.93945, "lr_stage": "cosine_decay", "mfu": 57.707771, "stage_name": "permutation_cycle", "status": "training", "step": 908, "tokens_per_sec_per_gpu": 100459.433985, "vram_gb": 20.709735} {"adam_lr": 6.63608949489552e-05, "dt_ms": 162.475348, "global_tokens_per_sec": 100839.913563, "grad_norm": 23.25, "loss": 3.410407, "lr_stage": "cosine_decay", "mfu": 57.926333, "stage_name": "permutation_cycle", "status": "training", "step": 909, "tokens_per_sec_per_gpu": 100839.913563, "vram_gb": 20.709735} {"adam_lr": 6.627840772285784e-05, "dt_ms": 163.818598, "global_tokens_per_sec": 100013.064577, "grad_norm": 30.125, "loss": 3.573805, "lr_stage": "cosine_decay", "mfu": 57.451359, "stage_name": "permutation_cycle", "status": "training", "step": 910, "tokens_per_sec_per_gpu": 100013.064577, "vram_gb": 20.709735} {"adam_lr": 6.619587090990748e-05, "dt_ms": 162.382841, "global_tokens_per_sec": 100897.36014, "grad_norm": 19.375, "loss": 2.688976, "lr_stage": "cosine_decay", "mfu": 57.959332, "stage_name": "permutation_cycle", "status": "training", "step": 911, "tokens_per_sec_per_gpu": 100897.36014, "vram_gb": 20.709735} {"adam_lr": 6.611328476152557e-05, "dt_ms": 163.382769, "global_tokens_per_sec": 100279.852871, "grad_norm": 33.25, "loss": 3.060626, "lr_stage": "cosine_decay", "mfu": 57.604613, "stage_name": "permutation_cycle", "status": "training", "step": 912, "tokens_per_sec_per_gpu": 100279.852871, "vram_gb": 20.709735} {"adam_lr": 6.603064952928382e-05, "dt_ms": 164.156199, "global_tokens_per_sec": 99807.379493, "grad_norm": 47.75, "loss": 2.621933, "lr_stage": "cosine_decay", "mfu": 57.333206, "stage_name": "permutation_cycle", "status": "training", "step": 913, "tokens_per_sec_per_gpu": 99807.379493, "vram_gb": 20.709735} {"adam_lr": 6.594796546490351e-05, "dt_ms": 164.185524, "global_tokens_per_sec": 99789.552709, "grad_norm": 24.5, "loss": 2.999396, "lr_stage": "cosine_decay", "mfu": 57.322965, "stage_name": "permutation_cycle", "status": "training", "step": 914, "tokens_per_sec_per_gpu": 99789.552709, "vram_gb": 20.709735} {"adam_lr": 6.586523282025462e-05, "dt_ms": 162.038565, "global_tokens_per_sec": 101111.732458, "grad_norm": 22.375, "loss": 3.079353, "lr_stage": "cosine_decay", "mfu": 58.082476, "stage_name": "permutation_cycle", "status": "training", "step": 915, "tokens_per_sec_per_gpu": 101111.732458, "vram_gb": 20.709735} {"adam_lr": 6.578245184735513e-05, "dt_ms": 163.202286, "global_tokens_per_sec": 100390.750798, "grad_norm": 24.5, "loss": 2.610739, "lr_stage": "cosine_decay", "mfu": 57.668317, "stage_name": "permutation_cycle", "status": "training", "step": 916, "tokens_per_sec_per_gpu": 100390.750798, "vram_gb": 20.709735} {"adam_lr": 6.569962279837026e-05, "dt_ms": 163.920164, "global_tokens_per_sec": 99951.095639, "grad_norm": 42.5, "loss": 3.535545, "lr_stage": "cosine_decay", "mfu": 57.415762, "stage_name": "permutation_cycle", "status": "training", "step": 917, "tokens_per_sec_per_gpu": 99951.095639, "vram_gb": 20.709735} {"adam_lr": 6.561674592561163e-05, "dt_ms": 163.677692, "global_tokens_per_sec": 100099.162925, "grad_norm": 28.5, "loss": 2.922575, "lr_stage": "cosine_decay", "mfu": 57.500817, "stage_name": "permutation_cycle", "status": "training", "step": 918, "tokens_per_sec_per_gpu": 100099.162925, "vram_gb": 20.709735} {"adam_lr": 6.55338214815366e-05, "dt_ms": 163.198709, "global_tokens_per_sec": 100392.950725, "grad_norm": 33.0, "loss": 3.165391, "lr_stage": "cosine_decay", "mfu": 57.66958, "stage_name": "permutation_cycle", "status": "training", "step": 919, "tokens_per_sec_per_gpu": 100392.950725, "vram_gb": 20.709735} {"adam_lr": 6.545084971874738e-05, "dt_ms": 161.748648, "global_tokens_per_sec": 101292.964325, "grad_norm": 37.25, "loss": 3.483795, "lr_stage": "cosine_decay", "mfu": 58.186583, "stage_name": "permutation_cycle", "status": "training", "step": 920, "tokens_per_sec_per_gpu": 101292.964325, "vram_gb": 20.709735} {"adam_lr": 6.536783088999037e-05, "dt_ms": 164.336205, "global_tokens_per_sec": 99698.055258, "grad_norm": 26.375, "loss": 3.38057, "lr_stage": "cosine_decay", "mfu": 57.270406, "stage_name": "permutation_cycle", "status": "training", "step": 921, "tokens_per_sec_per_gpu": 99698.055258, "vram_gb": 20.709735} {"adam_lr": 6.528476524815528e-05, "dt_ms": 162.786484, "global_tokens_per_sec": 100647.176726, "grad_norm": 41.5, "loss": 3.667477, "lr_stage": "cosine_decay", "mfu": 57.815617, "stage_name": "permutation_cycle", "status": "training", "step": 922, "tokens_per_sec_per_gpu": 100647.176726, "vram_gb": 20.709735} {"adam_lr": 6.520165304627452e-05, "dt_ms": 162.620068, "global_tokens_per_sec": 100750.173347, "grad_norm": 25.75, "loss": 3.460994, "lr_stage": "cosine_decay", "mfu": 57.874783, "stage_name": "permutation_cycle", "status": "training", "step": 923, "tokens_per_sec_per_gpu": 100750.173347, "vram_gb": 20.709735} {"adam_lr": 6.511849453752223e-05, "dt_ms": 163.924456, "global_tokens_per_sec": 99948.478924, "grad_norm": 25.875, "loss": 3.60849, "lr_stage": "cosine_decay", "mfu": 57.414259, "stage_name": "permutation_cycle", "status": "training", "step": 924, "tokens_per_sec_per_gpu": 99948.478924, "vram_gb": 20.709735} {"adam_lr": 6.503528997521366e-05, "dt_ms": 164.045811, "global_tokens_per_sec": 99874.540716, "grad_norm": 28.875, "loss": 2.289711, "lr_stage": "cosine_decay", "mfu": 57.371786, "stage_name": "permutation_cycle", "status": "training", "step": 925, "tokens_per_sec_per_gpu": 99874.540716, "vram_gb": 20.709735} {"adam_lr": 6.495203961280434e-05, "dt_ms": 162.536621, "global_tokens_per_sec": 100801.898611, "grad_norm": 33.75, "loss": 2.840216, "lr_stage": "cosine_decay", "mfu": 57.904496, "stage_name": "permutation_cycle", "status": "training", "step": 926, "tokens_per_sec_per_gpu": 100801.898611, "vram_gb": 20.709735} {"adam_lr": 6.48687437038893e-05, "dt_ms": 163.05089, "global_tokens_per_sec": 100483.965485, "grad_norm": 33.0, "loss": 2.742934, "lr_stage": "cosine_decay", "mfu": 57.721863, "stage_name": "permutation_cycle", "status": "training", "step": 927, "tokens_per_sec_per_gpu": 100483.965485, "vram_gb": 20.709735} {"adam_lr": 6.478540250220234e-05, "dt_ms": 163.970232, "global_tokens_per_sec": 99920.575821, "grad_norm": 46.5, "loss": 3.499438, "lr_stage": "cosine_decay", "mfu": 57.39823, "stage_name": "permutation_cycle", "status": "training", "step": 928, "tokens_per_sec_per_gpu": 99920.575821, "vram_gb": 20.709735} {"adam_lr": 6.47020162616152e-05, "dt_ms": 161.59153, "global_tokens_per_sec": 101391.452978, "grad_norm": 50.0, "loss": 2.546096, "lr_stage": "cosine_decay", "mfu": 58.243158, "stage_name": "permutation_cycle", "status": "training", "step": 929, "tokens_per_sec_per_gpu": 101391.452978, "vram_gb": 20.709735} {"adam_lr": 6.461858523613684e-05, "dt_ms": 163.936377, "global_tokens_per_sec": 99941.210991, "grad_norm": 31.125, "loss": 2.896765, "lr_stage": "cosine_decay", "mfu": 57.410084, "stage_name": "permutation_cycle", "status": "training", "step": 930, "tokens_per_sec_per_gpu": 99941.210991, "vram_gb": 20.709735} {"adam_lr": 6.453510967991263e-05, "dt_ms": 164.155245, "global_tokens_per_sec": 99807.959333, "grad_norm": 30.5, "loss": 3.412994, "lr_stage": "cosine_decay", "mfu": 57.333539, "stage_name": "permutation_cycle", "status": "training", "step": 931, "tokens_per_sec_per_gpu": 99807.959333, "vram_gb": 20.709735} {"adam_lr": 6.445158984722358e-05, "dt_ms": 162.902117, "global_tokens_per_sec": 100575.734216, "grad_norm": 30.75, "loss": 2.336355, "lr_stage": "cosine_decay", "mfu": 57.774578, "stage_name": "permutation_cycle", "status": "training", "step": 932, "tokens_per_sec_per_gpu": 100575.734216, "vram_gb": 20.709735} {"adam_lr": 6.436802599248562e-05, "dt_ms": 162.163496, "global_tokens_per_sec": 101033.835619, "grad_norm": 24.25, "loss": 3.04135, "lr_stage": "cosine_decay", "mfu": 58.037729, "stage_name": "permutation_cycle", "status": "training", "step": 933, "tokens_per_sec_per_gpu": 101033.835619, "vram_gb": 20.709735} {"adam_lr": 6.428441837024868e-05, "dt_ms": 163.128376, "global_tokens_per_sec": 100436.235565, "grad_norm": 27.25, "loss": 2.268137, "lr_stage": "cosine_decay", "mfu": 57.694445, "stage_name": "permutation_cycle", "status": "training", "step": 934, "tokens_per_sec_per_gpu": 100436.235565, "vram_gb": 20.709735} {"adam_lr": 6.420076723519614e-05, "dt_ms": 162.88209, "global_tokens_per_sec": 100588.100501, "grad_norm": 37.25, "loss": 3.026037, "lr_stage": "cosine_decay", "mfu": 57.781682, "stage_name": "permutation_cycle", "status": "training", "step": 935, "tokens_per_sec_per_gpu": 100588.100501, "vram_gb": 20.709735} {"adam_lr": 6.411707284214384e-05, "dt_ms": 162.021875, "global_tokens_per_sec": 101122.14762, "grad_norm": 33.25, "loss": 1.898861, "lr_stage": "cosine_decay", "mfu": 58.088459, "stage_name": "permutation_cycle", "status": "training", "step": 936, "tokens_per_sec_per_gpu": 101122.14762, "vram_gb": 20.709735} {"adam_lr": 6.40333354460394e-05, "dt_ms": 162.328959, "global_tokens_per_sec": 100930.851465, "grad_norm": 30.125, "loss": 2.980448, "lr_stage": "cosine_decay", "mfu": 57.978571, "stage_name": "permutation_cycle", "status": "training", "step": 937, "tokens_per_sec_per_gpu": 100930.851465, "vram_gb": 20.709735} {"adam_lr": 6.394955530196147e-05, "dt_ms": 163.09762, "global_tokens_per_sec": 100455.175244, "grad_norm": 23.625, "loss": 2.323677, "lr_stage": "cosine_decay", "mfu": 57.705324, "stage_name": "permutation_cycle", "status": "training", "step": 938, "tokens_per_sec_per_gpu": 100455.175244, "vram_gb": 20.709735} {"adam_lr": 6.386573266511891e-05, "dt_ms": 163.560152, "global_tokens_per_sec": 100171.097876, "grad_norm": 34.0, "loss": 3.251136, "lr_stage": "cosine_decay", "mfu": 57.542139, "stage_name": "permutation_cycle", "status": "training", "step": 939, "tokens_per_sec_per_gpu": 100171.097876, "vram_gb": 20.709735} {"adam_lr": 6.378186779084995e-05, "dt_ms": 161.72576, "global_tokens_per_sec": 101307.299777, "grad_norm": 25.25, "loss": 2.731187, "lr_stage": "cosine_decay", "mfu": 58.194818, "stage_name": "permutation_cycle", "status": "training", "step": 940, "tokens_per_sec_per_gpu": 101307.299777, "vram_gb": 20.709735} {"adam_lr": 6.369796093462164e-05, "dt_ms": 164.770842, "global_tokens_per_sec": 99435.068979, "grad_norm": 29.625, "loss": 2.826598, "lr_stage": "cosine_decay", "mfu": 57.119336, "stage_name": "permutation_cycle", "status": "training", "step": 941, "tokens_per_sec_per_gpu": 99435.068979, "vram_gb": 20.709735} {"adam_lr": 6.361401235202872e-05, "dt_ms": 163.073778, "global_tokens_per_sec": 100469.862081, "grad_norm": 30.625, "loss": 2.594223, "lr_stage": "cosine_decay", "mfu": 57.713761, "stage_name": "permutation_cycle", "status": "training", "step": 942, "tokens_per_sec_per_gpu": 100469.862081, "vram_gb": 20.709735} {"adam_lr": 6.353002229879318e-05, "dt_ms": 162.641764, "global_tokens_per_sec": 100736.733472, "grad_norm": 24.25, "loss": 2.653131, "lr_stage": "cosine_decay", "mfu": 57.867062, "stage_name": "permutation_cycle", "status": "training", "step": 943, "tokens_per_sec_per_gpu": 100736.733472, "vram_gb": 20.709735} {"adam_lr": 6.344599103076329e-05, "dt_ms": 162.32729, "global_tokens_per_sec": 100931.889162, "grad_norm": 18.375, "loss": 1.995175, "lr_stage": "cosine_decay", "mfu": 57.979167, "stage_name": "permutation_cycle", "status": "training", "step": 944, "tokens_per_sec_per_gpu": 100931.889162, "vram_gb": 20.709735} {"adam_lr": 6.336191880391284e-05, "dt_ms": 163.846493, "global_tokens_per_sec": 99996.037286, "grad_norm": 35.5, "loss": 2.966007, "lr_stage": "cosine_decay", "mfu": 57.441578, "stage_name": "permutation_cycle", "status": "training", "step": 945, "tokens_per_sec_per_gpu": 99996.037286, "vram_gb": 20.709735} {"adam_lr": 6.327780587434044e-05, "dt_ms": 164.663315, "global_tokens_per_sec": 99500.001066, "grad_norm": 29.375, "loss": 3.353224, "lr_stage": "cosine_decay", "mfu": 57.156636, "stage_name": "permutation_cycle", "status": "training", "step": 946, "tokens_per_sec_per_gpu": 99500.001066, "vram_gb": 20.709735} {"adam_lr": 6.319365249826865e-05, "dt_ms": 164.688587, "global_tokens_per_sec": 99484.732243, "grad_norm": 26.25, "loss": 3.032397, "lr_stage": "cosine_decay", "mfu": 57.147865, "stage_name": "permutation_cycle", "status": "training", "step": 947, "tokens_per_sec_per_gpu": 99484.732243, "vram_gb": 20.709735} {"adam_lr": 6.310945893204324e-05, "dt_ms": 164.465904, "global_tokens_per_sec": 99619.432223, "grad_norm": 36.0, "loss": 3.090741, "lr_stage": "cosine_decay", "mfu": 57.225241, "stage_name": "permutation_cycle", "status": "training", "step": 948, "tokens_per_sec_per_gpu": 99619.432223, "vram_gb": 20.709735} {"adam_lr": 6.302522543213243e-05, "dt_ms": 162.556648, "global_tokens_per_sec": 100789.479704, "grad_norm": 26.0, "loss": 2.60677, "lr_stage": "cosine_decay", "mfu": 57.897362, "stage_name": "permutation_cycle", "status": "training", "step": 949, "tokens_per_sec_per_gpu": 100789.479704, "vram_gb": 20.709735} {"adam_lr": 6.294095225512603e-05, "dt_ms": 161.661863, "global_tokens_per_sec": 101347.341066, "grad_norm": 38.75, "loss": 2.128408, "lr_stage": "cosine_decay", "mfu": 58.217819, "stage_name": "permutation_cycle", "status": "training", "step": 950, "tokens_per_sec_per_gpu": 101347.341066, "vram_gb": 20.709735} {"adam_lr": 6.285663965773482e-05, "dt_ms": 163.514137, "global_tokens_per_sec": 100199.287192, "grad_norm": 33.75, "loss": 2.229327, "lr_stage": "cosine_decay", "mfu": 57.558332, "stage_name": "permutation_cycle", "status": "training", "step": 951, "tokens_per_sec_per_gpu": 100199.287192, "vram_gb": 20.709735} {"adam_lr": 6.277228789678953e-05, "dt_ms": 163.576126, "global_tokens_per_sec": 100161.315656, "grad_norm": 27.25, "loss": 2.942996, "lr_stage": "cosine_decay", "mfu": 57.53652, "stage_name": "permutation_cycle", "status": "training", "step": 952, "tokens_per_sec_per_gpu": 100161.315656, "vram_gb": 20.709735} {"adam_lr": 6.268789722924029e-05, "dt_ms": 165.322781, "global_tokens_per_sec": 99103.099643, "grad_norm": 25.75, "loss": 3.131581, "lr_stage": "cosine_decay", "mfu": 56.92864, "stage_name": "permutation_cycle", "status": "training", "step": 953, "tokens_per_sec_per_gpu": 99103.099643, "vram_gb": 20.709735} {"adam_lr": 6.26034679121557e-05, "dt_ms": 163.936853, "global_tokens_per_sec": 99940.920295, "grad_norm": 32.25, "loss": 3.798313, "lr_stage": "cosine_decay", "mfu": 57.409917, "stage_name": "permutation_cycle", "status": "training", "step": 954, "tokens_per_sec_per_gpu": 99940.920295, "vram_gb": 20.709735} {"adam_lr": 6.251900020272208e-05, "dt_ms": 163.556337, "global_tokens_per_sec": 100173.434211, "grad_norm": 27.25, "loss": 3.309877, "lr_stage": "cosine_decay", "mfu": 57.543482, "stage_name": "permutation_cycle", "status": "training", "step": 955, "tokens_per_sec_per_gpu": 100173.434211, "vram_gb": 20.709735} {"adam_lr": 6.243449435824276e-05, "dt_ms": 163.377285, "global_tokens_per_sec": 100283.218684, "grad_norm": 31.125, "loss": 3.017041, "lr_stage": "cosine_decay", "mfu": 57.606546, "stage_name": "permutation_cycle", "status": "training", "step": 956, "tokens_per_sec_per_gpu": 100283.218684, "vram_gb": 20.709735} {"adam_lr": 6.234995063613716e-05, "dt_ms": 162.974119, "global_tokens_per_sec": 100531.299582, "grad_norm": 26.375, "loss": 2.938725, "lr_stage": "cosine_decay", "mfu": 57.749053, "stage_name": "permutation_cycle", "status": "training", "step": 957, "tokens_per_sec_per_gpu": 100531.299582, "vram_gb": 20.709735} {"adam_lr": 6.226536929394013e-05, "dt_ms": 162.291288, "global_tokens_per_sec": 100954.278963, "grad_norm": 35.0, "loss": 2.398865, "lr_stage": "cosine_decay", "mfu": 57.992029, "stage_name": "permutation_cycle", "status": "training", "step": 958, "tokens_per_sec_per_gpu": 100954.278963, "vram_gb": 20.709735} {"adam_lr": 6.218075058930113e-05, "dt_ms": 163.954973, "global_tokens_per_sec": 99929.875125, "grad_norm": 30.75, "loss": 2.854146, "lr_stage": "cosine_decay", "mfu": 57.403572, "stage_name": "permutation_cycle", "status": "training", "step": 959, "tokens_per_sec_per_gpu": 99929.875125, "vram_gb": 20.709735} {"adam_lr": 6.209609477998338e-05, "dt_ms": 163.170815, "global_tokens_per_sec": 100410.113468, "grad_norm": 40.5, "loss": 2.157334, "lr_stage": "cosine_decay", "mfu": 57.679439, "stage_name": "permutation_cycle", "status": "training", "step": 960, "tokens_per_sec_per_gpu": 100410.113468, "vram_gb": 20.709735} {"adam_lr": 6.201140212386319e-05, "dt_ms": 161.669493, "global_tokens_per_sec": 101342.558353, "grad_norm": 31.375, "loss": 2.942972, "lr_stage": "cosine_decay", "mfu": 58.215071, "stage_name": "permutation_cycle", "status": "training", "step": 961, "tokens_per_sec_per_gpu": 101342.558353, "vram_gb": 20.709735} {"adam_lr": 6.192667287892905e-05, "dt_ms": 160.572767, "global_tokens_per_sec": 102034.736524, "grad_norm": 35.75, "loss": 2.430254, "lr_stage": "cosine_decay", "mfu": 58.612685, "stage_name": "permutation_cycle", "status": "training", "step": 962, "tokens_per_sec_per_gpu": 102034.736524, "vram_gb": 20.709735} {"adam_lr": 6.184190730328095e-05, "dt_ms": 163.966894, "global_tokens_per_sec": 99922.609896, "grad_norm": 29.5, "loss": 2.729356, "lr_stage": "cosine_decay", "mfu": 57.399398, "stage_name": "permutation_cycle", "status": "training", "step": 963, "tokens_per_sec_per_gpu": 99922.609896, "vram_gb": 20.709735} {"adam_lr": 6.17571056551295e-05, "dt_ms": 163.581371, "global_tokens_per_sec": 100158.104001, "grad_norm": 27.125, "loss": 1.739313, "lr_stage": "cosine_decay", "mfu": 57.534675, "stage_name": "permutation_cycle", "status": "training", "step": 964, "tokens_per_sec_per_gpu": 100158.104001, "vram_gb": 20.709735} {"adam_lr": 6.167226819279528e-05, "dt_ms": 162.414551, "global_tokens_per_sec": 100877.66103, "grad_norm": 36.25, "loss": 1.788378, "lr_stage": "cosine_decay", "mfu": 57.948017, "stage_name": "permutation_cycle", "status": "training", "step": 965, "tokens_per_sec_per_gpu": 100877.66103, "vram_gb": 20.709735} {"adam_lr": 6.158739517470786e-05, "dt_ms": 162.171841, "global_tokens_per_sec": 101028.636862, "grad_norm": 22.875, "loss": 2.280369, "lr_stage": "cosine_decay", "mfu": 58.034743, "stage_name": "permutation_cycle", "status": "training", "step": 966, "tokens_per_sec_per_gpu": 101028.636862, "vram_gb": 20.709735} {"adam_lr": 6.150248685940523e-05, "dt_ms": 164.297104, "global_tokens_per_sec": 99721.782143, "grad_norm": 37.0, "loss": 3.170802, "lr_stage": "cosine_decay", "mfu": 57.284035, "stage_name": "permutation_cycle", "status": "training", "step": 967, "tokens_per_sec_per_gpu": 99721.782143, "vram_gb": 20.709735} {"adam_lr": 6.141754350553279e-05, "dt_ms": 160.373449, "global_tokens_per_sec": 102161.54899, "grad_norm": 31.875, "loss": 3.092151, "lr_stage": "cosine_decay", "mfu": 58.685531, "stage_name": "permutation_cycle", "status": "training", "step": 968, "tokens_per_sec_per_gpu": 102161.54899, "vram_gb": 20.709735} {"adam_lr": 6.133256537184275e-05, "dt_ms": 163.120747, "global_tokens_per_sec": 100440.933114, "grad_norm": 29.0, "loss": 3.288941, "lr_stage": "cosine_decay", "mfu": 57.697143, "stage_name": "permutation_cycle", "status": "training", "step": 969, "tokens_per_sec_per_gpu": 100440.933114, "vram_gb": 20.709735} {"adam_lr": 6.124755271719325e-05, "dt_ms": 163.440943, "global_tokens_per_sec": 100244.159896, "grad_norm": 33.0, "loss": 3.056942, "lr_stage": "cosine_decay", "mfu": 57.584109, "stage_name": "permutation_cycle", "status": "training", "step": 970, "tokens_per_sec_per_gpu": 100244.159896, "vram_gb": 20.709735} {"adam_lr": 6.116250580054757e-05, "dt_ms": 163.345337, "global_tokens_per_sec": 100302.832695, "grad_norm": 46.25, "loss": 3.596819, "lr_stage": "cosine_decay", "mfu": 57.617813, "stage_name": "permutation_cycle", "status": "training", "step": 971, "tokens_per_sec_per_gpu": 100302.832695, "vram_gb": 20.709735} {"adam_lr": 6.107742488097338e-05, "dt_ms": 162.545204, "global_tokens_per_sec": 100796.575847, "grad_norm": 43.0, "loss": 2.966633, "lr_stage": "cosine_decay", "mfu": 57.901438, "stage_name": "permutation_cycle", "status": "training", "step": 972, "tokens_per_sec_per_gpu": 100796.575847, "vram_gb": 20.709735} {"adam_lr": 6.099231021764188e-05, "dt_ms": 164.292812, "global_tokens_per_sec": 99724.387001, "grad_norm": 26.25, "loss": 2.239126, "lr_stage": "cosine_decay", "mfu": 57.285531, "stage_name": "permutation_cycle", "status": "training", "step": 973, "tokens_per_sec_per_gpu": 99724.387001, "vram_gb": 20.709735} {"adam_lr": 6.090716206982714e-05, "dt_ms": 162.428141, "global_tokens_per_sec": 100869.220909, "grad_norm": 43.25, "loss": 2.720625, "lr_stage": "cosine_decay", "mfu": 57.943168, "stage_name": "permutation_cycle", "status": "training", "step": 974, "tokens_per_sec_per_gpu": 100869.220909, "vram_gb": 20.709735} {"adam_lr": 6.0821980696905146e-05, "dt_ms": 163.909912, "global_tokens_per_sec": 99957.347235, "grad_norm": 40.5, "loss": 3.466526, "lr_stage": "cosine_decay", "mfu": 57.419353, "stage_name": "permutation_cycle", "status": "training", "step": 975, "tokens_per_sec_per_gpu": 99957.347235, "vram_gb": 20.709735} {"adam_lr": 6.073676635835317e-05, "dt_ms": 163.873911, "global_tokens_per_sec": 99979.30671, "grad_norm": 43.25, "loss": 3.218565, "lr_stage": "cosine_decay", "mfu": 57.431967, "stage_name": "permutation_cycle", "status": "training", "step": 976, "tokens_per_sec_per_gpu": 99979.30671, "vram_gb": 20.709735} {"adam_lr": 6.0651519313748836e-05, "dt_ms": 164.407492, "global_tokens_per_sec": 99654.826141, "grad_norm": 21.875, "loss": 2.229856, "lr_stage": "cosine_decay", "mfu": 57.245573, "stage_name": "permutation_cycle", "status": "training", "step": 977, "tokens_per_sec_per_gpu": 99654.826141, "vram_gb": 20.709735} {"adam_lr": 6.056623982276944e-05, "dt_ms": 162.752628, "global_tokens_per_sec": 100668.113127, "grad_norm": 25.0, "loss": 1.857048, "lr_stage": "cosine_decay", "mfu": 57.827644, "stage_name": "permutation_cycle", "status": "training", "step": 978, "tokens_per_sec_per_gpu": 100668.113127, "vram_gb": 20.709735} {"adam_lr": 6.04809281451911e-05, "dt_ms": 161.670208, "global_tokens_per_sec": 101342.109997, "grad_norm": 37.0, "loss": 3.047464, "lr_stage": "cosine_decay", "mfu": 58.214814, "stage_name": "permutation_cycle", "status": "training", "step": 979, "tokens_per_sec_per_gpu": 101342.109997, "vram_gb": 20.709735} {"adam_lr": 6.0395584540887963e-05, "dt_ms": 162.311077, "global_tokens_per_sec": 100941.970757, "grad_norm": 26.0, "loss": 2.652622, "lr_stage": "cosine_decay", "mfu": 57.984959, "stage_name": "permutation_cycle", "status": "training", "step": 980, "tokens_per_sec_per_gpu": 100941.970757, "vram_gb": 20.709735} {"adam_lr": 6.031020926983149e-05, "dt_ms": 163.530111, "global_tokens_per_sec": 100189.499466, "grad_norm": 15.9375, "loss": 1.884738, "lr_stage": "cosine_decay", "mfu": 57.55271, "stage_name": "permutation_cycle", "status": "training", "step": 981, "tokens_per_sec_per_gpu": 100189.499466, "vram_gb": 20.709735} {"adam_lr": 6.0224802592089513e-05, "dt_ms": 162.716627, "global_tokens_per_sec": 100690.38604, "grad_norm": 34.75, "loss": 2.575323, "lr_stage": "cosine_decay", "mfu": 57.840439, "stage_name": "permutation_cycle", "status": "training", "step": 982, "tokens_per_sec_per_gpu": 100690.38604, "vram_gb": 20.709735} {"adam_lr": 6.0139364767825626e-05, "dt_ms": 163.397551, "global_tokens_per_sec": 100270.780936, "grad_norm": 34.5, "loss": 3.034989, "lr_stage": "cosine_decay", "mfu": 57.599401, "stage_name": "permutation_cycle", "status": "training", "step": 983, "tokens_per_sec_per_gpu": 100270.780936, "vram_gb": 20.709735} {"adam_lr": 6.005389605729824e-05, "dt_ms": 163.832426, "global_tokens_per_sec": 100004.622973, "grad_norm": 21.875, "loss": 2.46248, "lr_stage": "cosine_decay", "mfu": 57.44651, "stage_name": "permutation_cycle", "status": "training", "step": 984, "tokens_per_sec_per_gpu": 100004.622973, "vram_gb": 20.709735} {"adam_lr": 5.9968396720859864e-05, "dt_ms": 162.772655, "global_tokens_per_sec": 100655.727161, "grad_norm": 30.0, "loss": 3.187096, "lr_stage": "cosine_decay", "mfu": 57.820529, "stage_name": "permutation_cycle", "status": "training", "step": 985, "tokens_per_sec_per_gpu": 100655.727161, "vram_gb": 20.709735} {"adam_lr": 5.988286701895631e-05, "dt_ms": 164.538383, "global_tokens_per_sec": 99575.549808, "grad_norm": 33.25, "loss": 3.418587, "lr_stage": "cosine_decay", "mfu": 57.200034, "stage_name": "permutation_cycle", "status": "training", "step": 986, "tokens_per_sec_per_gpu": 99575.549808, "vram_gb": 20.709735} {"adam_lr": 5.9797307212125886e-05, "dt_ms": 162.335157, "global_tokens_per_sec": 100926.997349, "grad_norm": 27.875, "loss": 2.374919, "lr_stage": "cosine_decay", "mfu": 57.976357, "stage_name": "permutation_cycle", "status": "training", "step": 987, "tokens_per_sec_per_gpu": 100926.997349, "vram_gb": 20.709735} {"adam_lr": 5.97117175609986e-05, "dt_ms": 163.436413, "global_tokens_per_sec": 100246.938355, "grad_norm": 39.0, "loss": 3.127371, "lr_stage": "cosine_decay", "mfu": 57.585705, "stage_name": "permutation_cycle", "status": "training", "step": 988, "tokens_per_sec_per_gpu": 100246.938355, "vram_gb": 20.709735} {"adam_lr": 5.9626098326295376e-05, "dt_ms": 162.410259, "global_tokens_per_sec": 100880.326625, "grad_norm": 32.25, "loss": 2.844599, "lr_stage": "cosine_decay", "mfu": 57.949548, "stage_name": "permutation_cycle", "status": "training", "step": 989, "tokens_per_sec_per_gpu": 100880.326625, "vram_gb": 20.709735} {"adam_lr": 5.9540449768827246e-05, "dt_ms": 162.074566, "global_tokens_per_sec": 101089.27277, "grad_norm": 31.125, "loss": 2.46947, "lr_stage": "cosine_decay", "mfu": 58.069574, "stage_name": "permutation_cycle", "status": "training", "step": 990, "tokens_per_sec_per_gpu": 101089.27277, "vram_gb": 20.709735} {"adam_lr": 5.945477214949456e-05, "dt_ms": 162.897587, "global_tokens_per_sec": 100578.531086, "grad_norm": 21.75, "loss": 2.760311, "lr_stage": "cosine_decay", "mfu": 57.776185, "stage_name": "permutation_cycle", "status": "training", "step": 991, "tokens_per_sec_per_gpu": 100578.531086, "vram_gb": 20.709735} {"adam_lr": 5.9369065729286245e-05, "dt_ms": 164.196968, "global_tokens_per_sec": 99782.597643, "grad_norm": 27.125, "loss": 2.391185, "lr_stage": "cosine_decay", "mfu": 57.31897, "stage_name": "permutation_cycle", "status": "training", "step": 992, "tokens_per_sec_per_gpu": 99782.597643, "vram_gb": 20.709735} {"adam_lr": 5.928333076927888e-05, "dt_ms": 160.777807, "global_tokens_per_sec": 101904.611598, "grad_norm": 37.25, "loss": 2.4837, "lr_stage": "cosine_decay", "mfu": 58.537937, "stage_name": "permutation_cycle", "status": "training", "step": 993, "tokens_per_sec_per_gpu": 101904.611598, "vram_gb": 20.709735} {"adam_lr": 5.9197567530636014e-05, "dt_ms": 165.127993, "global_tokens_per_sec": 99220.003459, "grad_norm": 23.5, "loss": 2.542062, "lr_stage": "cosine_decay", "mfu": 56.995794, "stage_name": "permutation_cycle", "status": "training", "step": 994, "tokens_per_sec_per_gpu": 99220.003459, "vram_gb": 20.709735} {"adam_lr": 5.911177627460739e-05, "dt_ms": 165.302992, "global_tokens_per_sec": 99114.963468, "grad_norm": 30.75, "loss": 2.956184, "lr_stage": "cosine_decay", "mfu": 56.935455, "stage_name": "permutation_cycle", "status": "training", "step": 995, "tokens_per_sec_per_gpu": 99114.963468, "vram_gb": 20.709735} {"adam_lr": 5.902595726252801e-05, "dt_ms": 162.412882, "global_tokens_per_sec": 100878.697633, "grad_norm": 33.5, "loss": 2.319397, "lr_stage": "cosine_decay", "mfu": 57.948612, "stage_name": "permutation_cycle", "status": "training", "step": 996, "tokens_per_sec_per_gpu": 100878.697633, "vram_gb": 20.709735} {"adam_lr": 5.8940110755817476e-05, "dt_ms": 162.367582, "global_tokens_per_sec": 100906.842153, "grad_norm": 32.25, "loss": 2.583634, "lr_stage": "cosine_decay", "mfu": 57.964779, "stage_name": "permutation_cycle", "status": "training", "step": 997, "tokens_per_sec_per_gpu": 100906.842153, "vram_gb": 20.709735} {"adam_lr": 5.885423701597917e-05, "dt_ms": 162.293434, "global_tokens_per_sec": 100952.944193, "grad_norm": 25.25, "loss": 2.07159, "lr_stage": "cosine_decay", "mfu": 57.991262, "stage_name": "permutation_cycle", "status": "training", "step": 998, "tokens_per_sec_per_gpu": 100952.944193, "vram_gb": 20.709735} {"adam_lr": 5.876833630459936e-05, "dt_ms": 163.556337, "global_tokens_per_sec": 100173.434211, "grad_norm": 23.0, "loss": 2.50661, "lr_stage": "cosine_decay", "mfu": 57.543482, "stage_name": "permutation_cycle", "status": "training", "step": 999, "tokens_per_sec_per_gpu": 100173.434211, "vram_gb": 20.709735} {"adam_lr": 5.868240888334653e-05, "dt_ms": 164.13641, "global_tokens_per_sec": 99819.412548, "grad_norm": 31.25, "loss": 2.613387, "lr_stage": "cosine_decay", "mfu": 57.340118, "stage_name": "permutation_cycle", "status": "training", "step": 1000, "tokens_per_sec_per_gpu": 99819.412548, "vram_gb": 20.709735} {"adam_lr": 5.859645501397048e-05, "dt_ms": 164.33239, "global_tokens_per_sec": 99700.369579, "grad_norm": 22.5, "loss": 2.038649, "lr_stage": "cosine_decay", "mfu": 57.271735, "stage_name": "permutation_cycle", "status": "training", "step": 1001, "tokens_per_sec_per_gpu": 99700.369579, "vram_gb": 20.709735} {"adam_lr": 5.851047495830163e-05, "dt_ms": 164.861441, "global_tokens_per_sec": 99380.424765, "grad_norm": 47.75, "loss": 3.388573, "lr_stage": "cosine_decay", "mfu": 57.087946, "stage_name": "permutation_cycle", "status": "training", "step": 1002, "tokens_per_sec_per_gpu": 99380.424765, "vram_gb": 20.709735} {"adam_lr": 5.842446897825014e-05, "dt_ms": 161.037445, "global_tokens_per_sec": 101740.31259, "grad_norm": 32.75, "loss": 1.635484, "lr_stage": "cosine_decay", "mfu": 58.443557, "stage_name": "permutation_cycle", "status": "training", "step": 1003, "tokens_per_sec_per_gpu": 101740.31259, "vram_gb": 20.709735} {"adam_lr": 5.833843733580512e-05, "dt_ms": 160.654783, "global_tokens_per_sec": 101982.646695, "grad_norm": 42.25, "loss": 3.150194, "lr_stage": "cosine_decay", "mfu": 58.582763, "stage_name": "permutation_cycle", "status": "training", "step": 1004, "tokens_per_sec_per_gpu": 101982.646695, "vram_gb": 20.709735} {"adam_lr": 5.8252380293033884e-05, "dt_ms": 164.31284, "global_tokens_per_sec": 99712.232161, "grad_norm": 25.75, "loss": 2.492452, "lr_stage": "cosine_decay", "mfu": 57.278549, "stage_name": "permutation_cycle", "status": "training", "step": 1005, "tokens_per_sec_per_gpu": 99712.232161, "vram_gb": 20.709735} {"adam_lr": 5.816629811208112e-05, "dt_ms": 164.751768, "global_tokens_per_sec": 99446.580682, "grad_norm": 36.25, "loss": 3.001836, "lr_stage": "cosine_decay", "mfu": 57.125949, "stage_name": "permutation_cycle", "status": "training", "step": 1006, "tokens_per_sec_per_gpu": 99446.580682, "vram_gb": 20.709735} {"adam_lr": 5.8080191055168064e-05, "dt_ms": 164.988995, "global_tokens_per_sec": 99303.593187, "grad_norm": 29.5, "loss": 3.541051, "lr_stage": "cosine_decay", "mfu": 57.043811, "stage_name": "permutation_cycle", "status": "training", "step": 1007, "tokens_per_sec_per_gpu": 99303.593187, "vram_gb": 20.709735} {"adam_lr": 5.799405938459175e-05, "dt_ms": 165.360451, "global_tokens_per_sec": 99080.523343, "grad_norm": 27.375, "loss": 2.798853, "lr_stage": "cosine_decay", "mfu": 56.915671, "stage_name": "permutation_cycle", "status": "training", "step": 1008, "tokens_per_sec_per_gpu": 99080.523343, "vram_gb": 20.709735} {"adam_lr": 5.790790336272419e-05, "dt_ms": 163.66148, "global_tokens_per_sec": 100109.078844, "grad_norm": 27.375, "loss": 3.088369, "lr_stage": "cosine_decay", "mfu": 57.506513, "stage_name": "permutation_cycle", "status": "training", "step": 1009, "tokens_per_sec_per_gpu": 100109.078844, "vram_gb": 20.709735} {"adam_lr": 5.782172325201155e-05, "dt_ms": 160.652637, "global_tokens_per_sec": 101984.008833, "grad_norm": 35.5, "loss": 3.123515, "lr_stage": "cosine_decay", "mfu": 58.583545, "stage_name": "permutation_cycle", "status": "training", "step": 1010, "tokens_per_sec_per_gpu": 101984.008833, "vram_gb": 20.709735} {"adam_lr": 5.773551931497342e-05, "dt_ms": 162.386179, "global_tokens_per_sec": 100895.286187, "grad_norm": 40.75, "loss": 2.466953, "lr_stage": "cosine_decay", "mfu": 57.958141, "stage_name": "permutation_cycle", "status": "training", "step": 1011, "tokens_per_sec_per_gpu": 100895.286187, "vram_gb": 20.709735} {"adam_lr": 5.764929181420191e-05, "dt_ms": 165.136337, "global_tokens_per_sec": 99214.989686, "grad_norm": 28.0, "loss": 3.25576, "lr_stage": "cosine_decay", "mfu": 56.992914, "stage_name": "permutation_cycle", "status": "training", "step": 1012, "tokens_per_sec_per_gpu": 99214.989686, "vram_gb": 20.709735} {"adam_lr": 5.756304101236097e-05, "dt_ms": 162.49156, "global_tokens_per_sec": 100829.852343, "grad_norm": 22.625, "loss": 3.097382, "lr_stage": "cosine_decay", "mfu": 57.920553, "stage_name": "permutation_cycle", "status": "training", "step": 1013, "tokens_per_sec_per_gpu": 100829.852343, "vram_gb": 20.709735} {"adam_lr": 5.747676717218549e-05, "dt_ms": 162.976027, "global_tokens_per_sec": 100530.123039, "grad_norm": 27.5, "loss": 1.957288, "lr_stage": "cosine_decay", "mfu": 57.748377, "stage_name": "permutation_cycle", "status": "training", "step": 1014, "tokens_per_sec_per_gpu": 100530.123039, "vram_gb": 20.709735} {"adam_lr": 5.7390470556480545e-05, "dt_ms": 165.062904, "global_tokens_per_sec": 99259.128293, "grad_norm": 43.0, "loss": 3.241955, "lr_stage": "cosine_decay", "mfu": 57.018269, "stage_name": "permutation_cycle", "status": "training", "step": 1015, "tokens_per_sec_per_gpu": 99259.128293, "vram_gb": 20.709735} {"adam_lr": 5.730415142812059e-05, "dt_ms": 162.240744, "global_tokens_per_sec": 100985.73042, "grad_norm": 39.5, "loss": 3.361473, "lr_stage": "cosine_decay", "mfu": 58.010096, "stage_name": "permutation_cycle", "status": "training", "step": 1016, "tokens_per_sec_per_gpu": 100985.73042, "vram_gb": 20.709735} {"adam_lr": 5.7217810050048656e-05, "dt_ms": 161.514521, "global_tokens_per_sec": 101439.795843, "grad_norm": 28.25, "loss": 2.836878, "lr_stage": "cosine_decay", "mfu": 58.270928, "stage_name": "permutation_cycle", "status": "training", "step": 1017, "tokens_per_sec_per_gpu": 101439.795843, "vram_gb": 20.709735} {"adam_lr": 5.713144668527559e-05, "dt_ms": 165.416002, "global_tokens_per_sec": 99047.249207, "grad_norm": 30.125, "loss": 3.013559, "lr_stage": "cosine_decay", "mfu": 56.896557, "stage_name": "permutation_cycle", "status": "training", "step": 1018, "tokens_per_sec_per_gpu": 99047.249207, "vram_gb": 20.709735} {"adam_lr": 5.7045061596879134e-05, "dt_ms": 164.818764, "global_tokens_per_sec": 99406.157581, "grad_norm": 33.25, "loss": 3.526068, "lr_stage": "cosine_decay", "mfu": 57.102728, "stage_name": "permutation_cycle", "status": "training", "step": 1019, "tokens_per_sec_per_gpu": 99406.157581, "vram_gb": 20.709735} {"adam_lr": 5.695865504800327e-05, "dt_ms": 161.309719, "global_tokens_per_sec": 101568.585531, "grad_norm": 23.375, "loss": 3.312799, "lr_stage": "cosine_decay", "mfu": 58.34491, "stage_name": "permutation_cycle", "status": "training", "step": 1020, "tokens_per_sec_per_gpu": 101568.585531, "vram_gb": 20.709735} {"adam_lr": 5.687222730185733e-05, "dt_ms": 162.315845, "global_tokens_per_sec": 100939.005373, "grad_norm": 40.25, "loss": 3.185869, "lr_stage": "cosine_decay", "mfu": 57.983255, "stage_name": "permutation_cycle", "status": "training", "step": 1021, "tokens_per_sec_per_gpu": 100939.005373, "vram_gb": 20.709735} {"adam_lr": 5.6785778621715225e-05, "dt_ms": 163.335562, "global_tokens_per_sec": 100308.83553, "grad_norm": 39.75, "loss": 3.630091, "lr_stage": "cosine_decay", "mfu": 57.621261, "stage_name": "permutation_cycle", "status": "training", "step": 1022, "tokens_per_sec_per_gpu": 100308.83553, "vram_gb": 20.709735} {"adam_lr": 5.669930927091461e-05, "dt_ms": 163.001537, "global_tokens_per_sec": 100514.389429, "grad_norm": 21.125, "loss": 2.632651, "lr_stage": "cosine_decay", "mfu": 57.739339, "stage_name": "permutation_cycle", "status": "training", "step": 1023, "tokens_per_sec_per_gpu": 100514.389429, "vram_gb": 20.709735} {"adam_lr": 5.661281951285613e-05, "dt_ms": 163.94639, "global_tokens_per_sec": 99935.106743, "grad_norm": 26.0, "loss": 2.233732, "lr_stage": "cosine_decay", "mfu": 57.406577, "stage_name": "permutation_cycle", "status": "training", "step": 1024, "tokens_per_sec_per_gpu": 99935.106743, "vram_gb": 20.709735} {"adam_lr": 5.6526309611002594e-05, "dt_ms": 164.083719, "global_tokens_per_sec": 99851.466523, "grad_norm": 36.75, "loss": 3.308736, "lr_stage": "cosine_decay", "mfu": 57.358531, "stage_name": "permutation_cycle", "status": "training", "step": 1025, "tokens_per_sec_per_gpu": 99851.466523, "vram_gb": 20.709735} {"adam_lr": 5.643977982887815e-05, "dt_ms": 162.62126, "global_tokens_per_sec": 100749.434799, "grad_norm": 34.25, "loss": 2.707768, "lr_stage": "cosine_decay", "mfu": 57.874358, "stage_name": "permutation_cycle", "status": "training", "step": 1026, "tokens_per_sec_per_gpu": 100749.434799, "vram_gb": 20.709735} {"adam_lr": 5.6353230430067525e-05, "dt_ms": 163.328409, "global_tokens_per_sec": 100313.228303, "grad_norm": 13.9375, "loss": 1.992392, "lr_stage": "cosine_decay", "mfu": 57.623785, "stage_name": "permutation_cycle", "status": "training", "step": 1027, "tokens_per_sec_per_gpu": 100313.228303, "vram_gb": 20.709735} {"adam_lr": 5.6266661678215216e-05, "dt_ms": 163.850069, "global_tokens_per_sec": 99993.85472, "grad_norm": 24.5, "loss": 2.17464, "lr_stage": "cosine_decay", "mfu": 57.440324, "stage_name": "permutation_cycle", "status": "training", "step": 1028, "tokens_per_sec_per_gpu": 99993.85472, "vram_gb": 20.709735} {"adam_lr": 5.618007383702464e-05, "dt_ms": 164.075851, "global_tokens_per_sec": 99856.25463, "grad_norm": 33.25, "loss": 2.857842, "lr_stage": "cosine_decay", "mfu": 57.361281, "stage_name": "permutation_cycle", "status": "training", "step": 1029, "tokens_per_sec_per_gpu": 99856.25463, "vram_gb": 20.709735} {"adam_lr": 5.6093467170257374e-05, "dt_ms": 163.793802, "global_tokens_per_sec": 100028.204815, "grad_norm": 24.125, "loss": 2.894069, "lr_stage": "cosine_decay", "mfu": 57.460056, "stage_name": "permutation_cycle", "status": "training", "step": 1030, "tokens_per_sec_per_gpu": 100028.204815, "vram_gb": 20.709735} {"adam_lr": 5.6006841941732355e-05, "dt_ms": 163.350821, "global_tokens_per_sec": 100299.465566, "grad_norm": 16.875, "loss": 2.462449, "lr_stage": "cosine_decay", "mfu": 57.615879, "stage_name": "permutation_cycle", "status": "training", "step": 1031, "tokens_per_sec_per_gpu": 100299.465566, "vram_gb": 20.709735} {"adam_lr": 5.5920198415325064e-05, "dt_ms": 164.24489, "global_tokens_per_sec": 99753.48383, "grad_norm": 37.75, "loss": 3.612669, "lr_stage": "cosine_decay", "mfu": 57.302246, "stage_name": "permutation_cycle", "status": "training", "step": 1032, "tokens_per_sec_per_gpu": 99753.48383, "vram_gb": 20.709735} {"adam_lr": 5.583353685496666e-05, "dt_ms": 162.396908, "global_tokens_per_sec": 100888.620487, "grad_norm": 36.0, "loss": 2.801448, "lr_stage": "cosine_decay", "mfu": 57.954312, "stage_name": "permutation_cycle", "status": "training", "step": 1033, "tokens_per_sec_per_gpu": 100888.620487, "vram_gb": 20.709735} {"adam_lr": 5.574685752464334e-05, "dt_ms": 161.275864, "global_tokens_per_sec": 101589.907066, "grad_norm": 39.25, "loss": 2.494519, "lr_stage": "cosine_decay", "mfu": 58.357158, "stage_name": "permutation_cycle", "status": "training", "step": 1034, "tokens_per_sec_per_gpu": 101589.907066, "vram_gb": 20.709735} {"adam_lr": 5.566016068839535e-05, "dt_ms": 163.409233, "global_tokens_per_sec": 100263.612342, "grad_norm": 31.125, "loss": 2.032189, "lr_stage": "cosine_decay", "mfu": 57.595283, "stage_name": "permutation_cycle", "status": "training", "step": 1035, "tokens_per_sec_per_gpu": 100263.612342, "vram_gb": 20.709735} {"adam_lr": 5.557344661031627e-05, "dt_ms": 160.817146, "global_tokens_per_sec": 101879.683708, "grad_norm": 28.125, "loss": 2.45244, "lr_stage": "cosine_decay", "mfu": 58.523617, "stage_name": "permutation_cycle", "status": "training", "step": 1036, "tokens_per_sec_per_gpu": 101879.683708, "vram_gb": 20.709735} {"adam_lr": 5.548671555455226e-05, "dt_ms": 163.915634, "global_tokens_per_sec": 99953.857876, "grad_norm": 26.625, "loss": 3.155662, "lr_stage": "cosine_decay", "mfu": 57.417348, "stage_name": "permutation_cycle", "status": "training", "step": 1037, "tokens_per_sec_per_gpu": 99953.857876, "vram_gb": 20.709735} {"adam_lr": 5.539996778530115e-05, "dt_ms": 162.598848, "global_tokens_per_sec": 100763.321309, "grad_norm": 26.625, "loss": 2.376436, "lr_stage": "cosine_decay", "mfu": 57.882335, "stage_name": "permutation_cycle", "status": "training", "step": 1038, "tokens_per_sec_per_gpu": 100763.321309, "vram_gb": 20.709735} {"adam_lr": 5.5313203566811666e-05, "dt_ms": 163.741589, "global_tokens_per_sec": 100060.101657, "grad_norm": 24.875, "loss": 3.213516, "lr_stage": "cosine_decay", "mfu": 57.478379, "stage_name": "permutation_cycle", "status": "training", "step": 1039, "tokens_per_sec_per_gpu": 100060.101657, "vram_gb": 20.709735} {"adam_lr": 5.522642316338268e-05, "dt_ms": 163.08856, "global_tokens_per_sec": 100460.755736, "grad_norm": 42.75, "loss": 3.910503, "lr_stage": "cosine_decay", "mfu": 57.70853, "stage_name": "permutation_cycle", "status": "training", "step": 1040, "tokens_per_sec_per_gpu": 100460.755736, "vram_gb": 20.709735} {"adam_lr": 5.5139626839362347e-05, "dt_ms": 165.826082, "global_tokens_per_sec": 98802.310105, "grad_norm": 23.0, "loss": 3.043529, "lr_stage": "cosine_decay", "mfu": 56.755855, "stage_name": "permutation_cycle", "status": "training", "step": 1041, "tokens_per_sec_per_gpu": 98802.310105, "vram_gb": 20.709735} {"adam_lr": 5.5052814859147315e-05, "dt_ms": 163.353205, "global_tokens_per_sec": 100298.001667, "grad_norm": 27.375, "loss": 2.970192, "lr_stage": "cosine_decay", "mfu": 57.615038, "stage_name": "permutation_cycle", "status": "training", "step": 1042, "tokens_per_sec_per_gpu": 100298.001667, "vram_gb": 20.709735} {"adam_lr": 5.4965987487181957e-05, "dt_ms": 165.744305, "global_tokens_per_sec": 98851.058767, "grad_norm": 28.875, "loss": 3.050209, "lr_stage": "cosine_decay", "mfu": 56.783858, "stage_name": "permutation_cycle", "status": "training", "step": 1043, "tokens_per_sec_per_gpu": 98851.058767, "vram_gb": 20.709735} {"adam_lr": 5.487914498795747e-05, "dt_ms": 163.115501, "global_tokens_per_sec": 100444.162934, "grad_norm": 18.25, "loss": 1.822863, "lr_stage": "cosine_decay", "mfu": 57.698999, "stage_name": "permutation_cycle", "status": "training", "step": 1044, "tokens_per_sec_per_gpu": 100444.162934, "vram_gb": 20.709735} {"adam_lr": 5.4792287626011204e-05, "dt_ms": 162.550688, "global_tokens_per_sec": 100793.175487, "grad_norm": 29.5, "loss": 2.226977, "lr_stage": "cosine_decay", "mfu": 57.899485, "stage_name": "permutation_cycle", "status": "training", "step": 1045, "tokens_per_sec_per_gpu": 100793.175487, "vram_gb": 20.709735} {"adam_lr": 5.470541566592573e-05, "dt_ms": 162.517786, "global_tokens_per_sec": 100813.581089, "grad_norm": 31.25, "loss": 2.86171, "lr_stage": "cosine_decay", "mfu": 57.911207, "stage_name": "permutation_cycle", "status": "training", "step": 1046, "tokens_per_sec_per_gpu": 100813.581089, "vram_gb": 20.709735} {"adam_lr": 5.461852937232809e-05, "dt_ms": 162.062168, "global_tokens_per_sec": 101097.006105, "grad_norm": 33.75, "loss": 2.251186, "lr_stage": "cosine_decay", "mfu": 58.074017, "stage_name": "permutation_cycle", "status": "training", "step": 1047, "tokens_per_sec_per_gpu": 101097.006105, "vram_gb": 20.709735} {"adam_lr": 5.453162900988902e-05, "dt_ms": 161.534309, "global_tokens_per_sec": 101427.368973, "grad_norm": 40.25, "loss": 2.314739, "lr_stage": "cosine_decay", "mfu": 58.26379, "stage_name": "permutation_cycle", "status": "training", "step": 1048, "tokens_per_sec_per_gpu": 101427.368973, "vram_gb": 20.709735} {"adam_lr": 5.4444714843322085e-05, "dt_ms": 161.899567, "global_tokens_per_sec": 101198.54141, "grad_norm": 32.5, "loss": 1.745496, "lr_stage": "cosine_decay", "mfu": 58.132343, "stage_name": "permutation_cycle", "status": "training", "step": 1049, "tokens_per_sec_per_gpu": 101198.54141, "vram_gb": 20.709735} {"adam_lr": 5.435778713738292e-05, "dt_ms": 164.496183, "global_tokens_per_sec": 99601.095064, "grad_norm": 35.0, "loss": 2.534042, "lr_stage": "cosine_decay", "mfu": 57.214708, "stage_name": "permutation_cycle", "status": "training", "step": 1050, "tokens_per_sec_per_gpu": 99601.095064, "vram_gb": 20.709735} {"adam_lr": 5.427084615686838e-05, "dt_ms": 165.253639, "global_tokens_per_sec": 99144.563939, "grad_norm": 34.25, "loss": 1.928169, "lr_stage": "cosine_decay", "mfu": 56.952459, "stage_name": "permutation_cycle", "status": "training", "step": 1051, "tokens_per_sec_per_gpu": 99144.563939, "vram_gb": 20.709735} {"adam_lr": 5.418389216661579e-05, "dt_ms": 162.348032, "global_tokens_per_sec": 100918.993587, "grad_norm": 39.0, "loss": 3.370962, "lr_stage": "cosine_decay", "mfu": 57.97176, "stage_name": "permutation_cycle", "status": "training", "step": 1052, "tokens_per_sec_per_gpu": 100918.993587, "vram_gb": 20.709735} {"adam_lr": 5.409692543150205e-05, "dt_ms": 163.352728, "global_tokens_per_sec": 100298.294443, "grad_norm": 25.5, "loss": 2.806508, "lr_stage": "cosine_decay", "mfu": 57.615206, "stage_name": "permutation_cycle", "status": "training", "step": 1053, "tokens_per_sec_per_gpu": 100298.294443, "vram_gb": 20.709735} {"adam_lr": 5.4009946216442944e-05, "dt_ms": 164.435863, "global_tokens_per_sec": 99637.631669, "grad_norm": 25.0, "loss": 2.763417, "lr_stage": "cosine_decay", "mfu": 57.235696, "stage_name": "permutation_cycle", "status": "training", "step": 1054, "tokens_per_sec_per_gpu": 99637.631669, "vram_gb": 20.709735} {"adam_lr": 5.392295478639225e-05, "dt_ms": 162.945986, "global_tokens_per_sec": 100548.656784, "grad_norm": 29.625, "loss": 2.671554, "lr_stage": "cosine_decay", "mfu": 57.759024, "stage_name": "permutation_cycle", "status": "training", "step": 1055, "tokens_per_sec_per_gpu": 100548.656784, "vram_gb": 20.709735} {"adam_lr": 5.383595140634093e-05, "dt_ms": 163.107157, "global_tokens_per_sec": 100449.301711, "grad_norm": 30.125, "loss": 2.29391, "lr_stage": "cosine_decay", "mfu": 57.70195, "stage_name": "permutation_cycle", "status": "training", "step": 1056, "tokens_per_sec_per_gpu": 100449.301711, "vram_gb": 20.709735} {"adam_lr": 5.374893634131639e-05, "dt_ms": 163.043737, "global_tokens_per_sec": 100488.373611, "grad_norm": 25.875, "loss": 2.716432, "lr_stage": "cosine_decay", "mfu": 57.724395, "stage_name": "permutation_cycle", "status": "training", "step": 1057, "tokens_per_sec_per_gpu": 100488.373611, "vram_gb": 20.709735} {"adam_lr": 5.366190985638159e-05, "dt_ms": 162.845373, "global_tokens_per_sec": 100610.779924, "grad_norm": 36.0, "loss": 2.882743, "lr_stage": "cosine_decay", "mfu": 57.79471, "stage_name": "permutation_cycle", "status": "training", "step": 1058, "tokens_per_sec_per_gpu": 100610.779924, "vram_gb": 20.709735} {"adam_lr": 5.35748722166343e-05, "dt_ms": 163.093328, "global_tokens_per_sec": 100457.818558, "grad_norm": 33.75, "loss": 2.737709, "lr_stage": "cosine_decay", "mfu": 57.706843, "stage_name": "permutation_cycle", "status": "training", "step": 1059, "tokens_per_sec_per_gpu": 100457.818558, "vram_gb": 20.709735} {"adam_lr": 5.348782368720626e-05, "dt_ms": 164.429188, "global_tokens_per_sec": 99641.676893, "grad_norm": 29.875, "loss": 2.751845, "lr_stage": "cosine_decay", "mfu": 57.23802, "stage_name": "permutation_cycle", "status": "training", "step": 1060, "tokens_per_sec_per_gpu": 99641.676893, "vram_gb": 20.709735} {"adam_lr": 5.340076453326241e-05, "dt_ms": 165.317774, "global_tokens_per_sec": 99106.101065, "grad_norm": 37.25, "loss": 2.78013, "lr_stage": "cosine_decay", "mfu": 56.930364, "stage_name": "permutation_cycle", "status": "training", "step": 1061, "tokens_per_sec_per_gpu": 99106.101065, "vram_gb": 20.709735} {"adam_lr": 5.3313695020000024e-05, "dt_ms": 162.196159, "global_tokens_per_sec": 101013.489249, "grad_norm": 28.375, "loss": 1.978195, "lr_stage": "cosine_decay", "mfu": 58.026042, "stage_name": "permutation_cycle", "status": "training", "step": 1062, "tokens_per_sec_per_gpu": 101013.489249, "vram_gb": 20.709735} {"adam_lr": 5.3226615412647907e-05, "dt_ms": 163.737535, "global_tokens_per_sec": 100062.578518, "grad_norm": 25.125, "loss": 2.318913, "lr_stage": "cosine_decay", "mfu": 57.479802, "stage_name": "permutation_cycle", "status": "training", "step": 1063, "tokens_per_sec_per_gpu": 100062.578518, "vram_gb": 20.709735} {"adam_lr": 5.313952597646568e-05, "dt_ms": 162.040472, "global_tokens_per_sec": 101110.54229, "grad_norm": 30.625, "loss": 2.678472, "lr_stage": "cosine_decay", "mfu": 58.081793, "stage_name": "permutation_cycle", "status": "training", "step": 1064, "tokens_per_sec_per_gpu": 101110.54229, "vram_gb": 20.709735} {"adam_lr": 5.3052426976742855e-05, "dt_ms": 163.300991, "global_tokens_per_sec": 100330.070833, "grad_norm": 32.0, "loss": 2.183495, "lr_stage": "cosine_decay", "mfu": 57.63346, "stage_name": "permutation_cycle", "status": "training", "step": 1065, "tokens_per_sec_per_gpu": 100330.070833, "vram_gb": 20.709735} {"adam_lr": 5.296531867879809e-05, "dt_ms": 162.459373, "global_tokens_per_sec": 100849.828789, "grad_norm": 22.125, "loss": 2.744254, "lr_stage": "cosine_decay", "mfu": 57.932029, "stage_name": "permutation_cycle", "status": "training", "step": 1066, "tokens_per_sec_per_gpu": 100849.828789, "vram_gb": 20.709735} {"adam_lr": 5.2878201347978374e-05, "dt_ms": 164.166689, "global_tokens_per_sec": 99801.001701, "grad_norm": 31.875, "loss": 2.611183, "lr_stage": "cosine_decay", "mfu": 57.329542, "stage_name": "permutation_cycle", "status": "training", "step": 1067, "tokens_per_sec_per_gpu": 99801.001701, "vram_gb": 20.709735} {"adam_lr": 5.279107524965819e-05, "dt_ms": 163.371325, "global_tokens_per_sec": 100286.877432, "grad_norm": 37.5, "loss": 3.286634, "lr_stage": "cosine_decay", "mfu": 57.608648, "stage_name": "permutation_cycle", "status": "training", "step": 1068, "tokens_per_sec_per_gpu": 100286.877432, "vram_gb": 20.709735} {"adam_lr": 5.270394064923878e-05, "dt_ms": 162.430286, "global_tokens_per_sec": 100867.888387, "grad_norm": 26.625, "loss": 2.232606, "lr_stage": "cosine_decay", "mfu": 57.942403, "stage_name": "permutation_cycle", "status": "training", "step": 1069, "tokens_per_sec_per_gpu": 100867.888387, "vram_gb": 20.709735} {"adam_lr": 5.26167978121472e-05, "dt_ms": 163.29813, "global_tokens_per_sec": 100331.82864, "grad_norm": 34.75, "loss": 2.658179, "lr_stage": "cosine_decay", "mfu": 57.634469, "stage_name": "permutation_cycle", "status": "training", "step": 1070, "tokens_per_sec_per_gpu": 100331.82864, "vram_gb": 20.709735} {"adam_lr": 5.252964700383567e-05, "dt_ms": 163.525343, "global_tokens_per_sec": 100192.420975, "grad_norm": 36.25, "loss": 2.579423, "lr_stage": "cosine_decay", "mfu": 57.554388, "stage_name": "permutation_cycle", "status": "training", "step": 1071, "tokens_per_sec_per_gpu": 100192.420975, "vram_gb": 20.709735} {"adam_lr": 5.244248848978067e-05, "dt_ms": 163.323402, "global_tokens_per_sec": 100316.303474, "grad_norm": 16.25, "loss": 3.037029, "lr_stage": "cosine_decay", "mfu": 57.625551, "stage_name": "permutation_cycle", "status": "training", "step": 1072, "tokens_per_sec_per_gpu": 100316.303474, "vram_gb": 20.709735} {"adam_lr": 5.235532253548213e-05, "dt_ms": 164.062023, "global_tokens_per_sec": 99864.671203, "grad_norm": 30.5, "loss": 2.723951, "lr_stage": "cosine_decay", "mfu": 57.366116, "stage_name": "permutation_cycle", "status": "training", "step": 1073, "tokens_per_sec_per_gpu": 99864.671203, "vram_gb": 20.709735} {"adam_lr": 5.226814940646269e-05, "dt_ms": 164.784431, "global_tokens_per_sec": 99426.868516, "grad_norm": 26.625, "loss": 2.672361, "lr_stage": "cosine_decay", "mfu": 57.114625, "stage_name": "permutation_cycle", "status": "training", "step": 1074, "tokens_per_sec_per_gpu": 99426.868516, "vram_gb": 20.709735} {"adam_lr": 5.218096936826681e-05, "dt_ms": 162.342548, "global_tokens_per_sec": 100922.402441, "grad_norm": 31.875, "loss": 2.621052, "lr_stage": "cosine_decay", "mfu": 57.973718, "stage_name": "permutation_cycle", "status": "training", "step": 1075, "tokens_per_sec_per_gpu": 100922.402441, "vram_gb": 20.709735} {"adam_lr": 5.209378268645998e-05, "dt_ms": 162.199736, "global_tokens_per_sec": 101011.262042, "grad_norm": 25.875, "loss": 1.713032, "lr_stage": "cosine_decay", "mfu": 58.024762, "stage_name": "permutation_cycle", "status": "training", "step": 1076, "tokens_per_sec_per_gpu": 101011.262042, "vram_gb": 20.709735} {"adam_lr": 5.200658962662799e-05, "dt_ms": 164.325953, "global_tokens_per_sec": 99704.275239, "grad_norm": 42.25, "loss": 3.39603, "lr_stage": "cosine_decay", "mfu": 57.273979, "stage_name": "permutation_cycle", "status": "training", "step": 1077, "tokens_per_sec_per_gpu": 99704.275239, "vram_gb": 20.709735} {"adam_lr": 5.191939045437601e-05, "dt_ms": 163.059235, "global_tokens_per_sec": 100478.823161, "grad_norm": 28.625, "loss": 2.458829, "lr_stage": "cosine_decay", "mfu": 57.718909, "stage_name": "permutation_cycle", "status": "training", "step": 1078, "tokens_per_sec_per_gpu": 100478.823161, "vram_gb": 20.709735} {"adam_lr": 5.183218543532782e-05, "dt_ms": 162.837744, "global_tokens_per_sec": 100615.493815, "grad_norm": 22.625, "loss": 1.890442, "lr_stage": "cosine_decay", "mfu": 57.797418, "stage_name": "permutation_cycle", "status": "training", "step": 1079, "tokens_per_sec_per_gpu": 100615.493815, "vram_gb": 20.709735} {"adam_lr": 5.174497483512506e-05, "dt_ms": 165.810823, "global_tokens_per_sec": 98811.402417, "grad_norm": 31.375, "loss": 2.816705, "lr_stage": "cosine_decay", "mfu": 56.761078, "stage_name": "permutation_cycle", "status": "training", "step": 1080, "tokens_per_sec_per_gpu": 98811.402417, "vram_gb": 20.709735} {"adam_lr": 5.1657758919426316e-05, "dt_ms": 164.440155, "global_tokens_per_sec": 99635.031341, "grad_norm": 22.625, "loss": 2.423375, "lr_stage": "cosine_decay", "mfu": 57.234202, "stage_name": "permutation_cycle", "status": "training", "step": 1081, "tokens_per_sec_per_gpu": 99635.031341, "vram_gb": 20.709735} {"adam_lr": 5.157053795390642e-05, "dt_ms": 161.15737, "global_tokens_per_sec": 101664.602986, "grad_norm": 33.25, "loss": 2.691998, "lr_stage": "cosine_decay", "mfu": 58.400066, "stage_name": "permutation_cycle", "status": "training", "step": 1082, "tokens_per_sec_per_gpu": 101664.602986, "vram_gb": 20.709735} {"adam_lr": 5.148331220425554e-05, "dt_ms": 162.420511, "global_tokens_per_sec": 100873.959048, "grad_norm": 22.625, "loss": 2.93253, "lr_stage": "cosine_decay", "mfu": 57.94589, "stage_name": "permutation_cycle", "status": "training", "step": 1083, "tokens_per_sec_per_gpu": 100873.959048, "vram_gb": 20.709735} {"adam_lr": 5.139608193617845e-05, "dt_ms": 165.400743, "global_tokens_per_sec": 99056.386657, "grad_norm": 27.875, "loss": 2.40028, "lr_stage": "cosine_decay", "mfu": 56.901806, "stage_name": "permutation_cycle", "status": "training", "step": 1084, "tokens_per_sec_per_gpu": 99056.386657, "vram_gb": 20.709735} {"adam_lr": 5.1308847415393666e-05, "dt_ms": 164.012671, "global_tokens_per_sec": 99894.721233, "grad_norm": 31.5, "loss": 3.271645, "lr_stage": "cosine_decay", "mfu": 57.383378, "stage_name": "permutation_cycle", "status": "training", "step": 1085, "tokens_per_sec_per_gpu": 99894.721233, "vram_gb": 20.709735} {"adam_lr": 5.1221608907632665e-05, "dt_ms": 162.297964, "global_tokens_per_sec": 100950.126462, "grad_norm": 23.125, "loss": 2.501165, "lr_stage": "cosine_decay", "mfu": 57.989643, "stage_name": "permutation_cycle", "status": "training", "step": 1086, "tokens_per_sec_per_gpu": 100950.126462, "vram_gb": 20.709735} {"adam_lr": 5.113436667863908e-05, "dt_ms": 164.832354, "global_tokens_per_sec": 99397.961887, "grad_norm": 26.375, "loss": 2.389035, "lr_stage": "cosine_decay", "mfu": 57.09802, "stage_name": "permutation_cycle", "status": "training", "step": 1087, "tokens_per_sec_per_gpu": 99397.961887, "vram_gb": 20.709735} {"adam_lr": 5.104712099416785e-05, "dt_ms": 163.714886, "global_tokens_per_sec": 100076.42206, "grad_norm": 29.5, "loss": 2.721692, "lr_stage": "cosine_decay", "mfu": 57.487754, "stage_name": "permutation_cycle", "status": "training", "step": 1088, "tokens_per_sec_per_gpu": 100076.42206, "vram_gb": 20.709735} {"adam_lr": 5.09598721199845e-05, "dt_ms": 162.303209, "global_tokens_per_sec": 100946.864022, "grad_norm": 35.75, "loss": 3.234405, "lr_stage": "cosine_decay", "mfu": 57.987769, "stage_name": "permutation_cycle", "status": "training", "step": 1089, "tokens_per_sec_per_gpu": 100946.864022, "vram_gb": 20.709735} {"adam_lr": 5.0872620321864185e-05, "dt_ms": 164.767742, "global_tokens_per_sec": 99436.939449, "grad_norm": 29.75, "loss": 3.07084, "lr_stage": "cosine_decay", "mfu": 57.120411, "stage_name": "permutation_cycle", "status": "training", "step": 1090, "tokens_per_sec_per_gpu": 99436.939449, "vram_gb": 20.709735} {"adam_lr": 5.078536586559104e-05, "dt_ms": 164.60824, "global_tokens_per_sec": 99533.291816, "grad_norm": 20.875, "loss": 1.88678, "lr_stage": "cosine_decay", "mfu": 57.175759, "stage_name": "permutation_cycle", "status": "training", "step": 1091, "tokens_per_sec_per_gpu": 99533.291816, "vram_gb": 20.709735} {"adam_lr": 5.0698109016957274e-05, "dt_ms": 163.472891, "global_tokens_per_sec": 100224.568823, "grad_norm": 31.0, "loss": 2.699845, "lr_stage": "cosine_decay", "mfu": 57.572855, "stage_name": "permutation_cycle", "status": "training", "step": 1092, "tokens_per_sec_per_gpu": 100224.568823, "vram_gb": 20.709735} {"adam_lr": 5.061085004176236e-05, "dt_ms": 164.048672, "global_tokens_per_sec": 99872.798895, "grad_norm": 35.0, "loss": 2.748301, "lr_stage": "cosine_decay", "mfu": 57.370785, "stage_name": "permutation_cycle", "status": "training", "step": 1093, "tokens_per_sec_per_gpu": 99872.798895, "vram_gb": 20.709735} {"adam_lr": 5.052358920581229e-05, "dt_ms": 163.939714, "global_tokens_per_sec": 99939.176159, "grad_norm": 22.625, "loss": 2.430946, "lr_stage": "cosine_decay", "mfu": 57.408915, "stage_name": "permutation_cycle", "status": "training", "step": 1094, "tokens_per_sec_per_gpu": 99939.176159, "vram_gb": 20.709735} {"adam_lr": 5.04363267749187e-05, "dt_ms": 162.380695, "global_tokens_per_sec": 100898.69344, "grad_norm": 27.625, "loss": 1.187364, "lr_stage": "cosine_decay", "mfu": 57.960098, "stage_name": "permutation_cycle", "status": "training", "step": 1095, "tokens_per_sec_per_gpu": 100898.69344, "vram_gb": 20.709735} {"adam_lr": 5.034906301489808e-05, "dt_ms": 162.065029, "global_tokens_per_sec": 101095.221384, "grad_norm": 34.5, "loss": 2.094766, "lr_stage": "cosine_decay", "mfu": 58.072992, "stage_name": "permutation_cycle", "status": "training", "step": 1096, "tokens_per_sec_per_gpu": 101095.221384, "vram_gb": 20.709735} {"adam_lr": 5.026179819157098e-05, "dt_ms": 165.367842, "global_tokens_per_sec": 99076.095023, "grad_norm": 35.75, "loss": 2.862572, "lr_stage": "cosine_decay", "mfu": 56.913128, "stage_name": "permutation_cycle", "status": "training", "step": 1097, "tokens_per_sec_per_gpu": 99076.095023, "vram_gb": 20.709735} {"adam_lr": 5.017453257076119e-05, "dt_ms": 163.085461, "global_tokens_per_sec": 100462.664994, "grad_norm": 31.375, "loss": 2.169211, "lr_stage": "cosine_decay", "mfu": 57.709627, "stage_name": "permutation_cycle", "status": "training", "step": 1098, "tokens_per_sec_per_gpu": 100462.664994, "vram_gb": 20.709735} {"adam_lr": 5.008726641829492e-05, "dt_ms": 162.185431, "global_tokens_per_sec": 101020.171459, "grad_norm": 43.0, "loss": 3.555338, "lr_stage": "cosine_decay", "mfu": 58.02988, "stage_name": "permutation_cycle", "status": "training", "step": 1099, "tokens_per_sec_per_gpu": 101020.171459, "vram_gb": 20.709735} {"adam_lr": 5e-05, "dt_ms": 163.429499, "global_tokens_per_sec": 100251.179457, "grad_norm": 21.875, "loss": 1.841006, "lr_stage": "cosine_decay", "mfu": 57.588141, "stage_name": "permutation_cycle", "status": "training", "step": 1100, "tokens_per_sec_per_gpu": 100251.179457, "vram_gb": 20.709735} {"adam_lr": 4.99127335817051e-05, "dt_ms": 164.652109, "global_tokens_per_sec": 99506.772704, "grad_norm": 29.875, "loss": 2.067618, "lr_stage": "cosine_decay", "mfu": 57.160525, "stage_name": "permutation_cycle", "status": "training", "step": 1101, "tokens_per_sec_per_gpu": 99506.772704, "vram_gb": 20.709735} {"adam_lr": 4.9825467429238834e-05, "dt_ms": 163.280725, "global_tokens_per_sec": 100342.523295, "grad_norm": 24.5, "loss": 2.325766, "lr_stage": "cosine_decay", "mfu": 57.640613, "stage_name": "permutation_cycle", "status": "training", "step": 1102, "tokens_per_sec_per_gpu": 100342.523295, "vram_gb": 20.709735} {"adam_lr": 4.973820180842902e-05, "dt_ms": 165.006876, "global_tokens_per_sec": 99292.831899, "grad_norm": 27.75, "loss": 2.701971, "lr_stage": "cosine_decay", "mfu": 57.03763, "stage_name": "permutation_cycle", "status": "training", "step": 1103, "tokens_per_sec_per_gpu": 99292.831899, "vram_gb": 20.709735} {"adam_lr": 4.965093698510193e-05, "dt_ms": 164.680481, "global_tokens_per_sec": 99489.62928, "grad_norm": 27.75, "loss": 2.501915, "lr_stage": "cosine_decay", "mfu": 57.150678, "stage_name": "permutation_cycle", "status": "training", "step": 1104, "tokens_per_sec_per_gpu": 99489.62928, "vram_gb": 20.709735} {"adam_lr": 4.9563673225081314e-05, "dt_ms": 164.642096, "global_tokens_per_sec": 99512.824735, "grad_norm": 30.25, "loss": 3.27254, "lr_stage": "cosine_decay", "mfu": 57.164002, "stage_name": "permutation_cycle", "status": "training", "step": 1105, "tokens_per_sec_per_gpu": 99512.824735, "vram_gb": 20.709735} {"adam_lr": 4.947641079418773e-05, "dt_ms": 163.826942, "global_tokens_per_sec": 100007.970335, "grad_norm": 39.0, "loss": 2.549785, "lr_stage": "cosine_decay", "mfu": 57.448433, "stage_name": "permutation_cycle", "status": "training", "step": 1106, "tokens_per_sec_per_gpu": 100007.970335, "vram_gb": 20.709735} {"adam_lr": 4.938914995823764e-05, "dt_ms": 157.999039, "global_tokens_per_sec": 103696.833444, "grad_norm": 33.25, "loss": 3.475644, "lr_stage": "cosine_decay", "mfu": 59.567458, "stage_name": "permutation_cycle", "status": "training", "step": 1107, "tokens_per_sec_per_gpu": 103696.833444, "vram_gb": 20.709735} {"adam_lr": 4.9301890983042744e-05, "dt_ms": 154.494524, "global_tokens_per_sec": 106049.066178, "grad_norm": 35.25, "loss": 3.366536, "lr_stage": "cosine_decay", "mfu": 60.918671, "stage_name": "permutation_cycle", "status": "training", "step": 1108, "tokens_per_sec_per_gpu": 106049.066178, "vram_gb": 20.709735} {"adam_lr": 4.921463413440898e-05, "dt_ms": 156.676531, "global_tokens_per_sec": 104572.13925, "grad_norm": 44.5, "loss": 3.387302, "lr_stage": "cosine_decay", "mfu": 60.070267, "stage_name": "permutation_cycle", "status": "training", "step": 1109, "tokens_per_sec_per_gpu": 104572.13925, "vram_gb": 20.709735} {"adam_lr": 4.912737967813583e-05, "dt_ms": 157.438278, "global_tokens_per_sec": 104066.178743, "grad_norm": 28.75, "loss": 2.881578, "lr_stage": "cosine_decay", "mfu": 59.779624, "stage_name": "permutation_cycle", "status": "training", "step": 1110, "tokens_per_sec_per_gpu": 104066.178743, "vram_gb": 20.709735} {"adam_lr": 4.904012788001552e-05, "dt_ms": 154.983044, "global_tokens_per_sec": 105714.790547, "grad_norm": 28.125, "loss": 1.795485, "lr_stage": "cosine_decay", "mfu": 60.72665, "stage_name": "permutation_cycle", "status": "training", "step": 1111, "tokens_per_sec_per_gpu": 105714.790547, "vram_gb": 20.709735} {"adam_lr": 4.895287900583216e-05, "dt_ms": 155.815363, "global_tokens_per_sec": 105150.093623, "grad_norm": 22.875, "loss": 2.664832, "lr_stage": "cosine_decay", "mfu": 60.402267, "stage_name": "permutation_cycle", "status": "training", "step": 1112, "tokens_per_sec_per_gpu": 105150.093623, "vram_gb": 20.709735} {"adam_lr": 4.886563332136093e-05, "dt_ms": 155.090809, "global_tokens_per_sec": 105641.334387, "grad_norm": 37.0, "loss": 2.270925, "lr_stage": "cosine_decay", "mfu": 60.684454, "stage_name": "permutation_cycle", "status": "training", "step": 1113, "tokens_per_sec_per_gpu": 105641.334387, "vram_gb": 20.709735} {"adam_lr": 4.8778391092367346e-05, "dt_ms": 155.324936, "global_tokens_per_sec": 105482.09728, "grad_norm": 26.375, "loss": 2.976724, "lr_stage": "cosine_decay", "mfu": 60.592982, "stage_name": "permutation_cycle", "status": "training", "step": 1114, "tokens_per_sec_per_gpu": 105482.09728, "vram_gb": 20.709735} {"adam_lr": 4.869115258460635e-05, "dt_ms": 155.835629, "global_tokens_per_sec": 105136.419423, "grad_norm": 46.0, "loss": 2.847082, "lr_stage": "cosine_decay", "mfu": 60.394412, "stage_name": "permutation_cycle", "status": "training", "step": 1115, "tokens_per_sec_per_gpu": 105136.419423, "vram_gb": 20.709735} {"adam_lr": 4.860391806382157e-05, "dt_ms": 153.878212, "global_tokens_per_sec": 106473.813217, "grad_norm": 27.75, "loss": 2.494119, "lr_stage": "cosine_decay", "mfu": 61.162662, "stage_name": "permutation_cycle", "status": "training", "step": 1116, "tokens_per_sec_per_gpu": 106473.813217, "vram_gb": 20.709735} {"adam_lr": 4.851668779574446e-05, "dt_ms": 156.958103, "global_tokens_per_sec": 104384.543825, "grad_norm": 44.5, "loss": 3.14855, "lr_stage": "cosine_decay", "mfu": 59.962505, "stage_name": "permutation_cycle", "status": "training", "step": 1117, "tokens_per_sec_per_gpu": 104384.543825, "vram_gb": 20.709735} {"adam_lr": 4.8429462046093585e-05, "dt_ms": 155.306339, "global_tokens_per_sec": 105494.727888, "grad_norm": 39.0, "loss": 2.515937, "lr_stage": "cosine_decay", "mfu": 60.600238, "stage_name": "permutation_cycle", "status": "training", "step": 1118, "tokens_per_sec_per_gpu": 105494.727888, "vram_gb": 20.709735} {"adam_lr": 4.8342241080573696e-05, "dt_ms": 154.438496, "global_tokens_per_sec": 106087.53946, "grad_norm": 17.5, "loss": 1.71809, "lr_stage": "cosine_decay", "mfu": 60.940772, "stage_name": "permutation_cycle", "status": "training", "step": 1119, "tokens_per_sec_per_gpu": 106087.53946, "vram_gb": 20.709735} {"adam_lr": 4.825502516487497e-05, "dt_ms": 156.53801, "global_tokens_per_sec": 104664.67561, "grad_norm": 33.0, "loss": 3.21761, "lr_stage": "cosine_decay", "mfu": 60.123424, "stage_name": "permutation_cycle", "status": "training", "step": 1120, "tokens_per_sec_per_gpu": 104664.67561, "vram_gb": 20.709735} {"adam_lr": 4.816781456467218e-05, "dt_ms": 160.422564, "global_tokens_per_sec": 102130.271685, "grad_norm": 23.125, "loss": 3.037516, "lr_stage": "cosine_decay", "mfu": 58.667564, "stage_name": "permutation_cycle", "status": "training", "step": 1121, "tokens_per_sec_per_gpu": 102130.271685, "vram_gb": 20.709735} {"adam_lr": 4.8080609545624004e-05, "dt_ms": 156.391382, "global_tokens_per_sec": 104762.805774, "grad_norm": 25.5, "loss": 3.01949, "lr_stage": "cosine_decay", "mfu": 60.179793, "stage_name": "permutation_cycle", "status": "training", "step": 1122, "tokens_per_sec_per_gpu": 104762.805774, "vram_gb": 20.709735} {"adam_lr": 4.799341037337202e-05, "dt_ms": 155.482292, "global_tokens_per_sec": 105375.343846, "grad_norm": 26.75, "loss": 3.072025, "lr_stage": "cosine_decay", "mfu": 60.531659, "stage_name": "permutation_cycle", "status": "training", "step": 1123, "tokens_per_sec_per_gpu": 105375.343846, "vram_gb": 20.709735} {"adam_lr": 4.790621731354003e-05, "dt_ms": 156.171083, "global_tokens_per_sec": 104910.586762, "grad_norm": 42.0, "loss": 2.891794, "lr_stage": "cosine_decay", "mfu": 60.264685, "stage_name": "permutation_cycle", "status": "training", "step": 1124, "tokens_per_sec_per_gpu": 104910.586762, "vram_gb": 20.709735} {"adam_lr": 4.781903063173321e-05, "dt_ms": 155.622244, "global_tokens_per_sec": 105280.579379, "grad_norm": 31.75, "loss": 3.143618, "lr_stage": "cosine_decay", "mfu": 60.477223, "stage_name": "permutation_cycle", "status": "training", "step": 1125, "tokens_per_sec_per_gpu": 105280.579379, "vram_gb": 20.709735} {"adam_lr": 4.773185059353732e-05, "dt_ms": 154.541254, "global_tokens_per_sec": 106016.999159, "grad_norm": 37.25, "loss": 2.811077, "lr_stage": "cosine_decay", "mfu": 60.90025, "stage_name": "permutation_cycle", "status": "training", "step": 1126, "tokens_per_sec_per_gpu": 106016.999159, "vram_gb": 20.709735} {"adam_lr": 4.7644677464517874e-05, "dt_ms": 158.577919, "global_tokens_per_sec": 103318.293635, "grad_norm": 27.875, "loss": 2.713949, "lr_stage": "cosine_decay", "mfu": 59.35001, "stage_name": "permutation_cycle", "status": "training", "step": 1127, "tokens_per_sec_per_gpu": 103318.293635, "vram_gb": 20.709735} {"adam_lr": 4.755751151021934e-05, "dt_ms": 156.549215, "global_tokens_per_sec": 104657.183793, "grad_norm": 35.25, "loss": 2.674888, "lr_stage": "cosine_decay", "mfu": 60.11912, "stage_name": "permutation_cycle", "status": "training", "step": 1128, "tokens_per_sec_per_gpu": 104657.183793, "vram_gb": 20.709735} {"adam_lr": 4.747035299616434e-05, "dt_ms": 154.725075, "global_tokens_per_sec": 105891.045938, "grad_norm": 29.625, "loss": 2.018919, "lr_stage": "cosine_decay", "mfu": 60.827898, "stage_name": "permutation_cycle", "status": "training", "step": 1129, "tokens_per_sec_per_gpu": 105891.045938, "vram_gb": 20.709735} {"adam_lr": 4.738320218785281e-05, "dt_ms": 156.023741, "global_tokens_per_sec": 105009.660192, "grad_norm": 21.875, "loss": 2.542889, "lr_stage": "cosine_decay", "mfu": 60.321596, "stage_name": "permutation_cycle", "status": "training", "step": 1130, "tokens_per_sec_per_gpu": 105009.660192, "vram_gb": 20.709735} {"adam_lr": 4.7296059350761226e-05, "dt_ms": 157.536507, "global_tokens_per_sec": 104001.290546, "grad_norm": 41.0, "loss": 3.302971, "lr_stage": "cosine_decay", "mfu": 59.74235, "stage_name": "permutation_cycle", "status": "training", "step": 1131, "tokens_per_sec_per_gpu": 104001.290546, "vram_gb": 20.709735} {"adam_lr": 4.720892475034181e-05, "dt_ms": 154.653072, "global_tokens_per_sec": 105940.346029, "grad_norm": 20.75, "loss": 2.96087, "lr_stage": "cosine_decay", "mfu": 60.856218, "stage_name": "permutation_cycle", "status": "training", "step": 1132, "tokens_per_sec_per_gpu": 105940.346029, "vram_gb": 20.709735} {"adam_lr": 4.7121798652021644e-05, "dt_ms": 155.318022, "global_tokens_per_sec": 105486.792922, "grad_norm": 30.0, "loss": 2.804655, "lr_stage": "cosine_decay", "mfu": 60.59568, "stage_name": "permutation_cycle", "status": "training", "step": 1133, "tokens_per_sec_per_gpu": 105486.792922, "vram_gb": 20.709735} {"adam_lr": 4.703468132120193e-05, "dt_ms": 155.114412, "global_tokens_per_sec": 105625.259164, "grad_norm": 24.125, "loss": 2.22029, "lr_stage": "cosine_decay", "mfu": 60.67522, "stage_name": "permutation_cycle", "status": "training", "step": 1134, "tokens_per_sec_per_gpu": 105625.259164, "vram_gb": 20.709735} {"adam_lr": 4.694757302325715e-05, "dt_ms": 154.056311, "global_tokens_per_sec": 106350.722865, "grad_norm": 36.75, "loss": 2.032153, "lr_stage": "cosine_decay", "mfu": 61.091954, "stage_name": "permutation_cycle", "status": "training", "step": 1135, "tokens_per_sec_per_gpu": 106350.722865, "vram_gb": 20.709735} {"adam_lr": 4.6860474023534335e-05, "dt_ms": 156.009674, "global_tokens_per_sec": 105019.128445, "grad_norm": 31.625, "loss": 2.481569, "lr_stage": "cosine_decay", "mfu": 60.327035, "stage_name": "permutation_cycle", "status": "training", "step": 1136, "tokens_per_sec_per_gpu": 105019.128445, "vram_gb": 20.709735} {"adam_lr": 4.677338458735211e-05, "dt_ms": 156.05855, "global_tokens_per_sec": 104986.237617, "grad_norm": 28.75, "loss": 2.613675, "lr_stage": "cosine_decay", "mfu": 60.308141, "stage_name": "permutation_cycle", "status": "training", "step": 1137, "tokens_per_sec_per_gpu": 104986.237617, "vram_gb": 20.709735} {"adam_lr": 4.668630498000001e-05, "dt_ms": 154.844761, "global_tokens_per_sec": 105809.198227, "grad_norm": 29.375, "loss": 2.832747, "lr_stage": "cosine_decay", "mfu": 60.780882, "stage_name": "permutation_cycle", "status": "training", "step": 1138, "tokens_per_sec_per_gpu": 105809.198227, "vram_gb": 20.709735} {"adam_lr": 4.659923546673761e-05, "dt_ms": 156.527758, "global_tokens_per_sec": 104671.530766, "grad_norm": 27.75, "loss": 2.206933, "lr_stage": "cosine_decay", "mfu": 60.127362, "stage_name": "permutation_cycle", "status": "training", "step": 1139, "tokens_per_sec_per_gpu": 104671.530766, "vram_gb": 20.709735} {"adam_lr": 4.6512176312793736e-05, "dt_ms": 154.390574, "global_tokens_per_sec": 106120.468552, "grad_norm": 32.75, "loss": 2.879308, "lr_stage": "cosine_decay", "mfu": 60.959687, "stage_name": "permutation_cycle", "status": "training", "step": 1140, "tokens_per_sec_per_gpu": 106120.468552, "vram_gb": 20.709735} {"adam_lr": 4.6425127783365707e-05, "dt_ms": 154.965401, "global_tokens_per_sec": 105726.826288, "grad_norm": 28.375, "loss": 2.010248, "lr_stage": "cosine_decay", "mfu": 60.733564, "stage_name": "permutation_cycle", "status": "training", "step": 1141, "tokens_per_sec_per_gpu": 105726.826288, "vram_gb": 20.709735} {"adam_lr": 4.633809014361843e-05, "dt_ms": 156.553745, "global_tokens_per_sec": 104654.15549, "grad_norm": 20.5, "loss": 2.388533, "lr_stage": "cosine_decay", "mfu": 60.117381, "stage_name": "permutation_cycle", "status": "training", "step": 1142, "tokens_per_sec_per_gpu": 104654.15549, "vram_gb": 20.709735} {"adam_lr": 4.625106365868363e-05, "dt_ms": 154.307604, "global_tokens_per_sec": 106177.528474, "grad_norm": 28.875, "loss": 2.699312, "lr_stage": "cosine_decay", "mfu": 60.992465, "stage_name": "permutation_cycle", "status": "training", "step": 1143, "tokens_per_sec_per_gpu": 106177.528474, "vram_gb": 20.709735} {"adam_lr": 4.616404859365907e-05, "dt_ms": 155.584812, "global_tokens_per_sec": 105305.90854, "grad_norm": 29.5, "loss": 2.090374, "lr_stage": "cosine_decay", "mfu": 60.491773, "stage_name": "permutation_cycle", "status": "training", "step": 1144, "tokens_per_sec_per_gpu": 105305.90854, "vram_gb": 20.709735} {"adam_lr": 4.607704521360776e-05, "dt_ms": 155.229807, "global_tokens_per_sec": 105546.739555, "grad_norm": 37.25, "loss": 3.243487, "lr_stage": "cosine_decay", "mfu": 60.630115, "stage_name": "permutation_cycle", "status": "training", "step": 1145, "tokens_per_sec_per_gpu": 105546.739555, "vram_gb": 20.709735} {"adam_lr": 4.599005378355706e-05, "dt_ms": 153.940201, "global_tokens_per_sec": 106430.93821, "grad_norm": 28.625, "loss": 2.384174, "lr_stage": "cosine_decay", "mfu": 61.138033, "stage_name": "permutation_cycle", "status": "training", "step": 1146, "tokens_per_sec_per_gpu": 106430.93821, "vram_gb": 20.709735} {"adam_lr": 4.590307456849796e-05, "dt_ms": 154.685497, "global_tokens_per_sec": 105918.138983, "grad_norm": 27.625, "loss": 2.628681, "lr_stage": "cosine_decay", "mfu": 60.843461, "stage_name": "permutation_cycle", "status": "training", "step": 1147, "tokens_per_sec_per_gpu": 105918.138983, "vram_gb": 20.709735} {"adam_lr": 4.5816107833384234e-05, "dt_ms": 159.854174, "global_tokens_per_sec": 102493.413996, "grad_norm": 49.75, "loss": 3.337233, "lr_stage": "cosine_decay", "mfu": 58.876167, "stage_name": "permutation_cycle", "status": "training", "step": 1148, "tokens_per_sec_per_gpu": 102493.413996, "vram_gb": 20.709735} {"adam_lr": 4.5729153843131636e-05, "dt_ms": 155.376434, "global_tokens_per_sec": 105447.13599, "grad_norm": 33.0, "loss": 2.815485, "lr_stage": "cosine_decay", "mfu": 60.572899, "stage_name": "permutation_cycle", "status": "training", "step": 1149, "tokens_per_sec_per_gpu": 105447.13599, "vram_gb": 20.709735} {"adam_lr": 4.564221286261709e-05, "dt_ms": 156.237364, "global_tokens_per_sec": 104866.080686, "grad_norm": 27.625, "loss": 2.24455, "lr_stage": "cosine_decay", "mfu": 60.239119, "stage_name": "permutation_cycle", "status": "training", "step": 1150, "tokens_per_sec_per_gpu": 104866.080686, "vram_gb": 20.709735} {"adam_lr": 4.555528515667793e-05, "dt_ms": 156.379223, "global_tokens_per_sec": 104770.95166, "grad_norm": 23.75, "loss": 2.203417, "lr_stage": "cosine_decay", "mfu": 60.184473, "stage_name": "permutation_cycle", "status": "training", "step": 1151, "tokens_per_sec_per_gpu": 104770.95166, "vram_gb": 20.709735} {"adam_lr": 4.5468370990111006e-05, "dt_ms": 154.983759, "global_tokens_per_sec": 105714.302669, "grad_norm": 37.25, "loss": 2.50198, "lr_stage": "cosine_decay", "mfu": 60.72637, "stage_name": "permutation_cycle", "status": "training", "step": 1152, "tokens_per_sec_per_gpu": 105714.302669, "vram_gb": 20.709735} {"adam_lr": 4.5381470627671916e-05, "dt_ms": 156.224251, "global_tokens_per_sec": 104874.882848, "grad_norm": 37.25, "loss": 3.539258, "lr_stage": "cosine_decay", "mfu": 60.244175, "stage_name": "permutation_cycle", "status": "training", "step": 1153, "tokens_per_sec_per_gpu": 104874.882848, "vram_gb": 20.709735} {"adam_lr": 4.529458433407429e-05, "dt_ms": 156.425476, "global_tokens_per_sec": 104739.972102, "grad_norm": 25.75, "loss": 2.623927, "lr_stage": "cosine_decay", "mfu": 60.166677, "stage_name": "permutation_cycle", "status": "training", "step": 1154, "tokens_per_sec_per_gpu": 104739.972102, "vram_gb": 20.709735} {"adam_lr": 4.52077123739888e-05, "dt_ms": 155.406952, "global_tokens_per_sec": 105426.429122, "grad_norm": 25.125, "loss": 2.404667, "lr_stage": "cosine_decay", "mfu": 60.561004, "stage_name": "permutation_cycle", "status": "training", "step": 1155, "tokens_per_sec_per_gpu": 105426.429122, "vram_gb": 20.709735} {"adam_lr": 4.512085501204253e-05, "dt_ms": 154.702663, "global_tokens_per_sec": 105906.38608, "grad_norm": 30.375, "loss": 2.156547, "lr_stage": "cosine_decay", "mfu": 60.83671, "stage_name": "permutation_cycle", "status": "training", "step": 1156, "tokens_per_sec_per_gpu": 105906.38608, "vram_gb": 20.709735} {"adam_lr": 4.503401251281806e-05, "dt_ms": 154.295444, "global_tokens_per_sec": 106185.895859, "grad_norm": 30.75, "loss": 1.531291, "lr_stage": "cosine_decay", "mfu": 60.997271, "stage_name": "permutation_cycle", "status": "training", "step": 1157, "tokens_per_sec_per_gpu": 106185.895859, "vram_gb": 20.709735} {"adam_lr": 4.494718514085268e-05, "dt_ms": 155.378819, "global_tokens_per_sec": 105445.517973, "grad_norm": 34.25, "loss": 2.377545, "lr_stage": "cosine_decay", "mfu": 60.57197, "stage_name": "permutation_cycle", "status": "training", "step": 1158, "tokens_per_sec_per_gpu": 105445.517973, "vram_gb": 20.709735} {"adam_lr": 4.4860373160637665e-05, "dt_ms": 156.464815, "global_tokens_per_sec": 104713.637921, "grad_norm": 22.0, "loss": 2.353446, "lr_stage": "cosine_decay", "mfu": 60.15155, "stage_name": "permutation_cycle", "status": "training", "step": 1159, "tokens_per_sec_per_gpu": 104713.637921, "vram_gb": 20.709735} {"adam_lr": 4.477357683661734e-05, "dt_ms": 154.580355, "global_tokens_per_sec": 105990.18247, "grad_norm": 30.5, "loss": 2.112488, "lr_stage": "cosine_decay", "mfu": 60.884846, "stage_name": "permutation_cycle", "status": "training", "step": 1160, "tokens_per_sec_per_gpu": 105990.18247, "vram_gb": 20.709735} {"adam_lr": 4.468679643318835e-05, "dt_ms": 154.969215, "global_tokens_per_sec": 105724.223733, "grad_norm": 37.25, "loss": 2.245128, "lr_stage": "cosine_decay", "mfu": 60.732069, "stage_name": "permutation_cycle", "status": "training", "step": 1161, "tokens_per_sec_per_gpu": 105724.223733, "vram_gb": 20.709735} {"adam_lr": 4.460003221469886e-05, "dt_ms": 154.081345, "global_tokens_per_sec": 106333.443819, "grad_norm": 26.625, "loss": 1.518038, "lr_stage": "cosine_decay", "mfu": 61.082028, "stage_name": "permutation_cycle", "status": "training", "step": 1162, "tokens_per_sec_per_gpu": 106333.443819, "vram_gb": 20.709735} {"adam_lr": 4.451328444544774e-05, "dt_ms": 155.877113, "global_tokens_per_sec": 105108.43862, "grad_norm": 41.25, "loss": 3.478806, "lr_stage": "cosine_decay", "mfu": 60.378338, "stage_name": "permutation_cycle", "status": "training", "step": 1163, "tokens_per_sec_per_gpu": 105108.43862, "vram_gb": 20.709735} {"adam_lr": 4.442655338968373e-05, "dt_ms": 155.711889, "global_tokens_per_sec": 105219.967962, "grad_norm": 43.75, "loss": 2.795711, "lr_stage": "cosine_decay", "mfu": 60.442405, "stage_name": "permutation_cycle", "status": "training", "step": 1164, "tokens_per_sec_per_gpu": 105219.967962, "vram_gb": 20.709735} {"adam_lr": 4.433983931160467e-05, "dt_ms": 156.615496, "global_tokens_per_sec": 104612.892413, "grad_norm": 34.25, "loss": 2.631926, "lr_stage": "cosine_decay", "mfu": 60.093677, "stage_name": "permutation_cycle", "status": "training", "step": 1165, "tokens_per_sec_per_gpu": 104612.892413, "vram_gb": 20.709735} {"adam_lr": 4.425314247535668e-05, "dt_ms": 155.865192, "global_tokens_per_sec": 105116.477556, "grad_norm": 47.5, "loss": 2.879836, "lr_stage": "cosine_decay", "mfu": 60.382956, "stage_name": "permutation_cycle", "status": "training", "step": 1166, "tokens_per_sec_per_gpu": 105116.477556, "vram_gb": 20.709735} {"adam_lr": 4.416646314503334e-05, "dt_ms": 157.217741, "global_tokens_per_sec": 104212.157575, "grad_norm": 20.75, "loss": 2.536286, "lr_stage": "cosine_decay", "mfu": 59.86348, "stage_name": "permutation_cycle", "status": "training", "step": 1167, "tokens_per_sec_per_gpu": 104212.157575, "vram_gb": 20.709735} {"adam_lr": 4.407980158467495e-05, "dt_ms": 156.195879, "global_tokens_per_sec": 104893.932585, "grad_norm": 26.375, "loss": 2.466106, "lr_stage": "cosine_decay", "mfu": 60.255118, "stage_name": "permutation_cycle", "status": "training", "step": 1168, "tokens_per_sec_per_gpu": 104893.932585, "vram_gb": 20.709735} {"adam_lr": 4.399315805826765e-05, "dt_ms": 156.054258, "global_tokens_per_sec": 104989.124767, "grad_norm": 26.375, "loss": 2.3445, "lr_stage": "cosine_decay", "mfu": 60.3098, "stage_name": "permutation_cycle", "status": "training", "step": 1169, "tokens_per_sec_per_gpu": 104989.124767, "vram_gb": 20.709735} {"adam_lr": 4.390653282974264e-05, "dt_ms": 154.355526, "global_tokens_per_sec": 106144.563967, "grad_norm": 33.0, "loss": 3.166687, "lr_stage": "cosine_decay", "mfu": 60.973529, "stage_name": "permutation_cycle", "status": "training", "step": 1170, "tokens_per_sec_per_gpu": 106144.563967, "vram_gb": 20.709735} {"adam_lr": 4.3819926162975375e-05, "dt_ms": 153.403521, "global_tokens_per_sec": 106803.285463, "grad_norm": 36.0, "loss": 2.529669, "lr_stage": "cosine_decay", "mfu": 61.351924, "stage_name": "permutation_cycle", "status": "training", "step": 1171, "tokens_per_sec_per_gpu": 106803.285463, "vram_gb": 20.709735} {"adam_lr": 4.373333832178478e-05, "dt_ms": 154.783964, "global_tokens_per_sec": 105850.758438, "grad_norm": 51.75, "loss": 3.025176, "lr_stage": "cosine_decay", "mfu": 60.804755, "stage_name": "permutation_cycle", "status": "training", "step": 1172, "tokens_per_sec_per_gpu": 105850.758438, "vram_gb": 20.709735} {"adam_lr": 4.364676956993248e-05, "dt_ms": 157.304525, "global_tokens_per_sec": 104154.664088, "grad_norm": 41.0, "loss": 3.50973, "lr_stage": "cosine_decay", "mfu": 59.830453, "stage_name": "permutation_cycle", "status": "training", "step": 1173, "tokens_per_sec_per_gpu": 104154.664088, "vram_gb": 20.709735} {"adam_lr": 4.356022017112187e-05, "dt_ms": 154.678583, "global_tokens_per_sec": 105922.873528, "grad_norm": 41.0, "loss": 2.391705, "lr_stage": "cosine_decay", "mfu": 60.846181, "stage_name": "permutation_cycle", "status": "training", "step": 1174, "tokens_per_sec_per_gpu": 105922.873528, "vram_gb": 20.709735} {"adam_lr": 4.347369038899744e-05, "dt_ms": 155.906439, "global_tokens_per_sec": 105088.668071, "grad_norm": 24.125, "loss": 2.786058, "lr_stage": "cosine_decay", "mfu": 60.366981, "stage_name": "permutation_cycle", "status": "training", "step": 1175, "tokens_per_sec_per_gpu": 105088.668071, "vram_gb": 20.709735} {"adam_lr": 4.3387180487143876e-05, "dt_ms": 154.735565, "global_tokens_per_sec": 105883.866972, "grad_norm": 36.5, "loss": 2.687181, "lr_stage": "cosine_decay", "mfu": 60.823774, "stage_name": "permutation_cycle", "status": "training", "step": 1176, "tokens_per_sec_per_gpu": 105883.866972, "vram_gb": 20.709735} {"adam_lr": 4.3300690729085395e-05, "dt_ms": 155.777454, "global_tokens_per_sec": 105175.681973, "grad_norm": 34.75, "loss": 2.407879, "lr_stage": "cosine_decay", "mfu": 60.416965, "stage_name": "permutation_cycle", "status": "training", "step": 1177, "tokens_per_sec_per_gpu": 105175.681973, "vram_gb": 20.709735} {"adam_lr": 4.321422137828479e-05, "dt_ms": 155.516624, "global_tokens_per_sec": 105352.080897, "grad_norm": 28.75, "loss": 2.559342, "lr_stage": "cosine_decay", "mfu": 60.518296, "stage_name": "permutation_cycle", "status": "training", "step": 1178, "tokens_per_sec_per_gpu": 105352.080897, "vram_gb": 20.709735} {"adam_lr": 4.3127772698142676e-05, "dt_ms": 155.176163, "global_tokens_per_sec": 105583.226913, "grad_norm": 39.25, "loss": 2.160562, "lr_stage": "cosine_decay", "mfu": 60.651075, "stage_name": "permutation_cycle", "status": "training", "step": 1179, "tokens_per_sec_per_gpu": 105583.226913, "vram_gb": 20.709735} {"adam_lr": 4.3041344951996746e-05, "dt_ms": 156.014681, "global_tokens_per_sec": 105015.758193, "grad_norm": 25.0, "loss": 2.665834, "lr_stage": "cosine_decay", "mfu": 60.325099, "stage_name": "permutation_cycle", "status": "training", "step": 1180, "tokens_per_sec_per_gpu": 105015.758193, "vram_gb": 20.709735} {"adam_lr": 4.295493840312087e-05, "dt_ms": 154.522419, "global_tokens_per_sec": 106029.921798, "grad_norm": 28.125, "loss": 1.545709, "lr_stage": "cosine_decay", "mfu": 60.907674, "stage_name": "permutation_cycle", "status": "training", "step": 1181, "tokens_per_sec_per_gpu": 106029.921798, "vram_gb": 20.709735} {"adam_lr": 4.2868553314724425e-05, "dt_ms": 155.127048, "global_tokens_per_sec": 105616.655246, "grad_norm": 40.25, "loss": 1.873842, "lr_stage": "cosine_decay", "mfu": 60.670277, "stage_name": "permutation_cycle", "status": "training", "step": 1182, "tokens_per_sec_per_gpu": 105616.655246, "vram_gb": 20.709735} {"adam_lr": 4.278218994995135e-05, "dt_ms": 156.341076, "global_tokens_per_sec": 104796.515605, "grad_norm": 36.5, "loss": 3.105796, "lr_stage": "cosine_decay", "mfu": 60.199158, "stage_name": "permutation_cycle", "status": "training", "step": 1183, "tokens_per_sec_per_gpu": 104796.515605, "vram_gb": 20.709735} {"adam_lr": 4.269584857187943e-05, "dt_ms": 155.109406, "global_tokens_per_sec": 105628.668651, "grad_norm": 35.0, "loss": 2.185875, "lr_stage": "cosine_decay", "mfu": 60.677178, "stage_name": "permutation_cycle", "status": "training", "step": 1184, "tokens_per_sec_per_gpu": 105628.668651, "vram_gb": 20.709735} {"adam_lr": 4.260952944351947e-05, "dt_ms": 158.009052, "global_tokens_per_sec": 103690.261817, "grad_norm": 26.875, "loss": 1.836163, "lr_stage": "cosine_decay", "mfu": 59.563683, "stage_name": "permutation_cycle", "status": "training", "step": 1185, "tokens_per_sec_per_gpu": 103690.261817, "vram_gb": 20.709735} {"adam_lr": 4.252323282781453e-05, "dt_ms": 155.505657, "global_tokens_per_sec": 105359.511001, "grad_norm": 33.0, "loss": 3.405575, "lr_stage": "cosine_decay", "mfu": 60.522564, "stage_name": "permutation_cycle", "status": "training", "step": 1186, "tokens_per_sec_per_gpu": 105359.511001, "vram_gb": 20.709735} {"adam_lr": 4.243695898763904e-05, "dt_ms": 155.751228, "global_tokens_per_sec": 105193.391894, "grad_norm": 33.25, "loss": 2.581186, "lr_stage": "cosine_decay", "mfu": 60.427139, "stage_name": "permutation_cycle", "status": "training", "step": 1187, "tokens_per_sec_per_gpu": 105193.391894, "vram_gb": 20.709735} {"adam_lr": 4.23507081857981e-05, "dt_ms": 155.996799, "global_tokens_per_sec": 105027.795799, "grad_norm": 25.625, "loss": 3.089287, "lr_stage": "cosine_decay", "mfu": 60.332014, "stage_name": "permutation_cycle", "status": "training", "step": 1188, "tokens_per_sec_per_gpu": 105027.795799, "vram_gb": 20.709735} {"adam_lr": 4.226448068502661e-05, "dt_ms": 154.406786, "global_tokens_per_sec": 106109.326074, "grad_norm": 38.0, "loss": 2.764217, "lr_stage": "cosine_decay", "mfu": 60.953287, "stage_name": "permutation_cycle", "status": "training", "step": 1189, "tokens_per_sec_per_gpu": 106109.326074, "vram_gb": 20.709735} {"adam_lr": 4.2178276747988446e-05, "dt_ms": 157.15909, "global_tokens_per_sec": 104251.049021, "grad_norm": 28.5, "loss": 2.85482, "lr_stage": "cosine_decay", "mfu": 59.885821, "stage_name": "permutation_cycle", "status": "training", "step": 1190, "tokens_per_sec_per_gpu": 104251.049021, "vram_gb": 20.709735} {"adam_lr": 4.209209663727582e-05, "dt_ms": 153.969049, "global_tokens_per_sec": 106410.996613, "grad_norm": 30.75, "loss": 2.620455, "lr_stage": "cosine_decay", "mfu": 61.126578, "stage_name": "permutation_cycle", "status": "training", "step": 1191, "tokens_per_sec_per_gpu": 106410.996613, "vram_gb": 20.709735} {"adam_lr": 4.2005940615408264e-05, "dt_ms": 155.641317, "global_tokens_per_sec": 105267.677485, "grad_norm": 37.75, "loss": 2.435067, "lr_stage": "cosine_decay", "mfu": 60.469811, "stage_name": "permutation_cycle", "status": "training", "step": 1192, "tokens_per_sec_per_gpu": 105267.677485, "vram_gb": 20.709735} {"adam_lr": 4.1919808944831954e-05, "dt_ms": 154.195309, "global_tokens_per_sec": 106254.853923, "grad_norm": 30.875, "loss": 2.589389, "lr_stage": "cosine_decay", "mfu": 61.036883, "stage_name": "permutation_cycle", "status": "training", "step": 1193, "tokens_per_sec_per_gpu": 106254.853923, "vram_gb": 20.709735} {"adam_lr": 4.1833701887918904e-05, "dt_ms": 153.499603, "global_tokens_per_sec": 106736.432218, "grad_norm": 28.0, "loss": 1.991883, "lr_stage": "cosine_decay", "mfu": 61.313521, "stage_name": "permutation_cycle", "status": "training", "step": 1194, "tokens_per_sec_per_gpu": 106736.432218, "vram_gb": 20.709735} {"adam_lr": 4.174761970696612e-05, "dt_ms": 155.615807, "global_tokens_per_sec": 105284.934481, "grad_norm": 26.25, "loss": 2.494961, "lr_stage": "cosine_decay", "mfu": 60.479724, "stage_name": "permutation_cycle", "status": "training", "step": 1195, "tokens_per_sec_per_gpu": 105284.934481, "vram_gb": 20.709735} {"adam_lr": 4.166156266419489e-05, "dt_ms": 154.970884, "global_tokens_per_sec": 105723.085156, "grad_norm": 25.75, "loss": 2.063876, "lr_stage": "cosine_decay", "mfu": 60.731415, "stage_name": "permutation_cycle", "status": "training", "step": 1196, "tokens_per_sec_per_gpu": 105723.085156, "vram_gb": 20.709735} {"adam_lr": 4.157553102174988e-05, "dt_ms": 154.604673, "global_tokens_per_sec": 105973.51064, "grad_norm": 29.625, "loss": 1.91515, "lr_stage": "cosine_decay", "mfu": 60.875269, "stage_name": "permutation_cycle", "status": "training", "step": 1197, "tokens_per_sec_per_gpu": 105973.51064, "vram_gb": 20.709735} {"adam_lr": 4.1489525041698387e-05, "dt_ms": 156.370401, "global_tokens_per_sec": 104776.862214, "grad_norm": 30.0, "loss": 2.865653, "lr_stage": "cosine_decay", "mfu": 60.187868, "stage_name": "permutation_cycle", "status": "training", "step": 1198, "tokens_per_sec_per_gpu": 104776.862214, "vram_gb": 20.709735} {"adam_lr": 4.140354498602952e-05, "dt_ms": 155.318499, "global_tokens_per_sec": 105486.469071, "grad_norm": 35.25, "loss": 2.681035, "lr_stage": "cosine_decay", "mfu": 60.595494, "stage_name": "permutation_cycle", "status": "training", "step": 1199, "tokens_per_sec_per_gpu": 105486.469071, "vram_gb": 20.709735} {"adam_lr": 4.131759111665349e-05, "dt_ms": 156.022787, "global_tokens_per_sec": 105010.302054, "grad_norm": 37.75, "loss": 2.406809, "lr_stage": "cosine_decay", "mfu": 60.321965, "stage_name": "permutation_cycle", "status": "training", "step": 1200, "tokens_per_sec_per_gpu": 105010.302054, "vram_gb": 20.709735} {"adam_lr": 4.123166369540065e-05, "dt_ms": 155.134201, "global_tokens_per_sec": 105611.785726, "grad_norm": 34.5, "loss": 3.20318, "lr_stage": "cosine_decay", "mfu": 60.66748, "stage_name": "permutation_cycle", "status": "training", "step": 1201, "tokens_per_sec_per_gpu": 105611.785726, "vram_gb": 20.709735} {"adam_lr": 4.114576298402084e-05, "dt_ms": 155.426025, "global_tokens_per_sec": 105413.491459, "grad_norm": 25.0, "loss": 2.623606, "lr_stage": "cosine_decay", "mfu": 60.553572, "stage_name": "permutation_cycle", "status": "training", "step": 1202, "tokens_per_sec_per_gpu": 105413.491459, "vram_gb": 20.709735} {"adam_lr": 4.105988924418253e-05, "dt_ms": 155.800104, "global_tokens_per_sec": 105160.391839, "grad_norm": 24.75, "loss": 2.48251, "lr_stage": "cosine_decay", "mfu": 60.408182, "stage_name": "permutation_cycle", "status": "training", "step": 1203, "tokens_per_sec_per_gpu": 105160.391839, "vram_gb": 20.709735} {"adam_lr": 4.0974042737472006e-05, "dt_ms": 155.676365, "global_tokens_per_sec": 105243.978498, "grad_norm": 29.25, "loss": 2.827429, "lr_stage": "cosine_decay", "mfu": 60.456198, "stage_name": "permutation_cycle", "status": "training", "step": 1204, "tokens_per_sec_per_gpu": 105243.978498, "vram_gb": 20.709735} {"adam_lr": 4.088822372539263e-05, "dt_ms": 155.691385, "global_tokens_per_sec": 105233.825055, "grad_norm": 36.25, "loss": 2.750843, "lr_stage": "cosine_decay", "mfu": 60.450365, "stage_name": "permutation_cycle", "status": "training", "step": 1205, "tokens_per_sec_per_gpu": 105233.825055, "vram_gb": 20.709735} {"adam_lr": 4.080243246936399e-05, "dt_ms": 154.207706, "global_tokens_per_sec": 106246.311401, "grad_norm": 31.875, "loss": 2.539165, "lr_stage": "cosine_decay", "mfu": 61.031976, "stage_name": "permutation_cycle", "status": "training", "step": 1206, "tokens_per_sec_per_gpu": 106246.311401, "vram_gb": 20.709735} {"adam_lr": 4.071666923072115e-05, "dt_ms": 159.362793, "global_tokens_per_sec": 102809.443125, "grad_norm": 33.5, "loss": 3.04343, "lr_stage": "cosine_decay", "mfu": 59.057707, "stage_name": "permutation_cycle", "status": "training", "step": 1207, "tokens_per_sec_per_gpu": 102809.443125, "vram_gb": 20.709735} {"adam_lr": 4.063093427071376e-05, "dt_ms": 161.529541, "global_tokens_per_sec": 101430.363121, "grad_norm": 24.5, "loss": 1.735967, "lr_stage": "cosine_decay", "mfu": 58.26551, "stage_name": "permutation_cycle", "status": "training", "step": 1208, "tokens_per_sec_per_gpu": 101430.363121, "vram_gb": 20.709735} {"adam_lr": 4.054522785050543e-05, "dt_ms": 158.279181, "global_tokens_per_sec": 103513.29812, "grad_norm": 30.0, "loss": 2.628224, "lr_stage": "cosine_decay", "mfu": 59.462028, "stage_name": "permutation_cycle", "status": "training", "step": 1209, "tokens_per_sec_per_gpu": 103513.29812, "vram_gb": 20.709735} {"adam_lr": 4.045955023117276e-05, "dt_ms": 158.42247, "global_tokens_per_sec": 103419.672666, "grad_norm": 37.25, "loss": 2.643201, "lr_stage": "cosine_decay", "mfu": 59.408246, "stage_name": "permutation_cycle", "status": "training", "step": 1210, "tokens_per_sec_per_gpu": 103419.672666, "vram_gb": 20.709735} {"adam_lr": 4.037390167370464e-05, "dt_ms": 173.209429, "global_tokens_per_sec": 94590.693559, "grad_norm": 30.625, "loss": 2.84929, "lr_stage": "cosine_decay", "mfu": 54.33654, "stage_name": "permutation_cycle", "status": "training", "step": 1211, "tokens_per_sec_per_gpu": 94590.693559, "vram_gb": 20.709735} {"adam_lr": 4.028828243900141e-05, "dt_ms": 162.831545, "global_tokens_per_sec": 100619.324176, "grad_norm": 22.625, "loss": 2.746234, "lr_stage": "cosine_decay", "mfu": 57.799618, "stage_name": "permutation_cycle", "status": "training", "step": 1212, "tokens_per_sec_per_gpu": 100619.324176, "vram_gb": 20.709735} {"adam_lr": 4.020269278787411e-05, "dt_ms": 157.242537, "global_tokens_per_sec": 104195.724389, "grad_norm": 24.625, "loss": 2.182255, "lr_stage": "cosine_decay", "mfu": 59.85404, "stage_name": "permutation_cycle", "status": "training", "step": 1213, "tokens_per_sec_per_gpu": 104195.724389, "vram_gb": 20.709735} {"adam_lr": 4.0117132981043693e-05, "dt_ms": 156.248093, "global_tokens_per_sec": 104858.880016, "grad_norm": 21.75, "loss": 2.930095, "lr_stage": "cosine_decay", "mfu": 60.234982, "stage_name": "permutation_cycle", "status": "training", "step": 1214, "tokens_per_sec_per_gpu": 104858.880016, "vram_gb": 20.709735} {"adam_lr": 4.003160327914015e-05, "dt_ms": 155.669928, "global_tokens_per_sec": 105248.330573, "grad_norm": 34.75, "loss": 2.669835, "lr_stage": "cosine_decay", "mfu": 60.458698, "stage_name": "permutation_cycle", "status": "training", "step": 1215, "tokens_per_sec_per_gpu": 105248.330573, "vram_gb": 20.709735} {"adam_lr": 3.9946103942701777e-05, "dt_ms": 159.043074, "global_tokens_per_sec": 103016.117732, "grad_norm": 23.375, "loss": 2.410008, "lr_stage": "cosine_decay", "mfu": 59.176429, "stage_name": "permutation_cycle", "status": "training", "step": 1216, "tokens_per_sec_per_gpu": 103016.117732, "vram_gb": 20.709735} {"adam_lr": 3.986063523217439e-05, "dt_ms": 156.907082, "global_tokens_per_sec": 104418.486613, "grad_norm": 29.5, "loss": 2.711502, "lr_stage": "cosine_decay", "mfu": 59.982003, "stage_name": "permutation_cycle", "status": "training", "step": 1217, "tokens_per_sec_per_gpu": 104418.486613, "vram_gb": 20.709735} {"adam_lr": 3.9775197407910485e-05, "dt_ms": 154.676914, "global_tokens_per_sec": 105924.016413, "grad_norm": 32.75, "loss": 2.750067, "lr_stage": "cosine_decay", "mfu": 60.846838, "stage_name": "permutation_cycle", "status": "training", "step": 1218, "tokens_per_sec_per_gpu": 105924.016413, "vram_gb": 20.709735} {"adam_lr": 3.968979073016852e-05, "dt_ms": 154.515266, "global_tokens_per_sec": 106034.829954, "grad_norm": 33.75, "loss": 2.512043, "lr_stage": "cosine_decay", "mfu": 60.910493, "stage_name": "permutation_cycle", "status": "training", "step": 1219, "tokens_per_sec_per_gpu": 106034.829954, "vram_gb": 20.709735} {"adam_lr": 3.960441545911204e-05, "dt_ms": 154.381275, "global_tokens_per_sec": 106126.860147, "grad_norm": 24.5, "loss": 2.320616, "lr_stage": "cosine_decay", "mfu": 60.963359, "stage_name": "permutation_cycle", "status": "training", "step": 1220, "tokens_per_sec_per_gpu": 106126.860147, "vram_gb": 20.709735} {"adam_lr": 3.951907185480892e-05, "dt_ms": 156.079292, "global_tokens_per_sec": 104972.285297, "grad_norm": 27.625, "loss": 2.557458, "lr_stage": "cosine_decay", "mfu": 60.300127, "stage_name": "permutation_cycle", "status": "training", "step": 1221, "tokens_per_sec_per_gpu": 104972.285297, "vram_gb": 20.709735} {"adam_lr": 3.943376017723057e-05, "dt_ms": 155.544519, "global_tokens_per_sec": 105333.187313, "grad_norm": 22.625, "loss": 2.200537, "lr_stage": "cosine_decay", "mfu": 60.507443, "stage_name": "permutation_cycle", "status": "training", "step": 1222, "tokens_per_sec_per_gpu": 105333.187313, "vram_gb": 20.709735} {"adam_lr": 3.9348480686251176e-05, "dt_ms": 158.39386, "global_tokens_per_sec": 103438.353066, "grad_norm": 18.375, "loss": 1.990291, "lr_stage": "cosine_decay", "mfu": 59.418977, "stage_name": "permutation_cycle", "status": "training", "step": 1223, "tokens_per_sec_per_gpu": 103438.353066, "vram_gb": 20.709735} {"adam_lr": 3.926323364164684e-05, "dt_ms": 155.740261, "global_tokens_per_sec": 105200.79963, "grad_norm": 23.0, "loss": 2.425106, "lr_stage": "cosine_decay", "mfu": 60.431394, "stage_name": "permutation_cycle", "status": "training", "step": 1224, "tokens_per_sec_per_gpu": 105200.79963, "vram_gb": 20.709735} {"adam_lr": 3.917801930309486e-05, "dt_ms": 156.23188, "global_tokens_per_sec": 104869.76141, "grad_norm": 18.75, "loss": 2.316584, "lr_stage": "cosine_decay", "mfu": 60.241233, "stage_name": "permutation_cycle", "status": "training", "step": 1225, "tokens_per_sec_per_gpu": 104869.76141, "vram_gb": 20.709735} {"adam_lr": 3.9092837930172884e-05, "dt_ms": 153.856516, "global_tokens_per_sec": 106488.827632, "grad_norm": 32.25, "loss": 1.969402, "lr_stage": "cosine_decay", "mfu": 61.171287, "stage_name": "permutation_cycle", "status": "training", "step": 1226, "tokens_per_sec_per_gpu": 106488.827632, "vram_gb": 20.709735} {"adam_lr": 3.900768978235812e-05, "dt_ms": 153.672934, "global_tokens_per_sec": 106616.042386, "grad_norm": 27.25, "loss": 2.806498, "lr_stage": "cosine_decay", "mfu": 61.244364, "stage_name": "permutation_cycle", "status": "training", "step": 1227, "tokens_per_sec_per_gpu": 106616.042386, "vram_gb": 20.709735} {"adam_lr": 3.892257511902664e-05, "dt_ms": 155.883789, "global_tokens_per_sec": 105103.937353, "grad_norm": 35.0, "loss": 2.092764, "lr_stage": "cosine_decay", "mfu": 60.375753, "stage_name": "permutation_cycle", "status": "training", "step": 1228, "tokens_per_sec_per_gpu": 105103.937353, "vram_gb": 20.709735} {"adam_lr": 3.883749419945244e-05, "dt_ms": 155.260563, "global_tokens_per_sec": 105525.831508, "grad_norm": 40.25, "loss": 3.412727, "lr_stage": "cosine_decay", "mfu": 60.618105, "stage_name": "permutation_cycle", "status": "training", "step": 1229, "tokens_per_sec_per_gpu": 105525.831508, "vram_gb": 20.709735} {"adam_lr": 3.875244728280676e-05, "dt_ms": 154.505253, "global_tokens_per_sec": 106041.702137, "grad_norm": 38.75, "loss": 2.81597, "lr_stage": "cosine_decay", "mfu": 60.914441, "stage_name": "permutation_cycle", "status": "training", "step": 1230, "tokens_per_sec_per_gpu": 106041.702137, "vram_gb": 20.709735} {"adam_lr": 3.8667434628157236e-05, "dt_ms": 156.838655, "global_tokens_per_sec": 104464.042686, "grad_norm": 41.0, "loss": 2.913954, "lr_stage": "cosine_decay", "mfu": 60.008172, "stage_name": "permutation_cycle", "status": "training", "step": 1231, "tokens_per_sec_per_gpu": 104464.042686, "vram_gb": 20.709735} {"adam_lr": 3.858245649446721e-05, "dt_ms": 154.593945, "global_tokens_per_sec": 105980.865213, "grad_norm": 24.0, "loss": 1.696831, "lr_stage": "cosine_decay", "mfu": 60.879494, "stage_name": "permutation_cycle", "status": "training", "step": 1232, "tokens_per_sec_per_gpu": 105980.865213, "vram_gb": 20.709735} {"adam_lr": 3.849751314059479e-05, "dt_ms": 153.246403, "global_tokens_per_sec": 106912.786904, "grad_norm": 36.25, "loss": 2.290799, "lr_stage": "cosine_decay", "mfu": 61.414825, "stage_name": "permutation_cycle", "status": "training", "step": 1233, "tokens_per_sec_per_gpu": 106912.786904, "vram_gb": 20.709735} {"adam_lr": 3.841260482529214e-05, "dt_ms": 156.147718, "global_tokens_per_sec": 104926.284961, "grad_norm": 19.125, "loss": 2.6589, "lr_stage": "cosine_decay", "mfu": 60.273702, "stage_name": "permutation_cycle", "status": "training", "step": 1234, "tokens_per_sec_per_gpu": 104926.284961, "vram_gb": 20.709735} {"adam_lr": 3.832773180720475e-05, "dt_ms": 154.849291, "global_tokens_per_sec": 105806.102891, "grad_norm": 21.375, "loss": 2.422911, "lr_stage": "cosine_decay", "mfu": 60.779104, "stage_name": "permutation_cycle", "status": "training", "step": 1235, "tokens_per_sec_per_gpu": 105806.102891, "vram_gb": 20.709735} {"adam_lr": 3.82428943448705e-05, "dt_ms": 153.160095, "global_tokens_per_sec": 106973.033524, "grad_norm": 31.75, "loss": 2.272449, "lr_stage": "cosine_decay", "mfu": 61.449433, "stage_name": "permutation_cycle", "status": "training", "step": 1236, "tokens_per_sec_per_gpu": 106973.033524, "vram_gb": 20.709735} {"adam_lr": 3.815809269671907e-05, "dt_ms": 155.709505, "global_tokens_per_sec": 105221.579065, "grad_norm": 27.25, "loss": 2.555894, "lr_stage": "cosine_decay", "mfu": 60.44333, "stage_name": "permutation_cycle", "status": "training", "step": 1237, "tokens_per_sec_per_gpu": 105221.579065, "vram_gb": 20.709735} {"adam_lr": 3.807332712107097e-05, "dt_ms": 155.239105, "global_tokens_per_sec": 105540.417643, "grad_norm": 27.5, "loss": 2.719711, "lr_stage": "cosine_decay", "mfu": 60.626484, "stage_name": "permutation_cycle", "status": "training", "step": 1238, "tokens_per_sec_per_gpu": 105540.417643, "vram_gb": 20.709735} {"adam_lr": 3.798859787613682e-05, "dt_ms": 155.601263, "global_tokens_per_sec": 105294.775115, "grad_norm": 26.5, "loss": 2.441863, "lr_stage": "cosine_decay", "mfu": 60.485377, "stage_name": "permutation_cycle", "status": "training", "step": 1239, "tokens_per_sec_per_gpu": 105294.775115, "vram_gb": 20.709735} {"adam_lr": 3.790390522001662e-05, "dt_ms": 153.696537, "global_tokens_per_sec": 106599.669179, "grad_norm": 27.125, "loss": 2.376222, "lr_stage": "cosine_decay", "mfu": 61.234959, "stage_name": "permutation_cycle", "status": "training", "step": 1240, "tokens_per_sec_per_gpu": 106599.669179, "vram_gb": 20.709735} {"adam_lr": 3.781924941069888e-05, "dt_ms": 156.349897, "global_tokens_per_sec": 104790.602834, "grad_norm": 28.125, "loss": 2.458564, "lr_stage": "cosine_decay", "mfu": 60.195761, "stage_name": "permutation_cycle", "status": "training", "step": 1241, "tokens_per_sec_per_gpu": 104790.602834, "vram_gb": 20.709735} {"adam_lr": 3.773463070605987e-05, "dt_ms": 154.896021, "global_tokens_per_sec": 105774.182616, "grad_norm": 44.25, "loss": 2.266679, "lr_stage": "cosine_decay", "mfu": 60.760767, "stage_name": "permutation_cycle", "status": "training", "step": 1242, "tokens_per_sec_per_gpu": 105774.182616, "vram_gb": 20.709735} {"adam_lr": 3.765004936386286e-05, "dt_ms": 154.823542, "global_tokens_per_sec": 105823.699848, "grad_norm": 29.0, "loss": 1.824652, "lr_stage": "cosine_decay", "mfu": 60.789212, "stage_name": "permutation_cycle", "status": "training", "step": 1243, "tokens_per_sec_per_gpu": 105823.699848, "vram_gb": 20.709735} {"adam_lr": 3.756550564175727e-05, "dt_ms": 155.281782, "global_tokens_per_sec": 105511.411404, "grad_norm": 21.875, "loss": 1.79424, "lr_stage": "cosine_decay", "mfu": 60.609821, "stage_name": "permutation_cycle", "status": "training", "step": 1244, "tokens_per_sec_per_gpu": 105511.411404, "vram_gb": 20.709735} {"adam_lr": 3.748099979727792e-05, "dt_ms": 154.308796, "global_tokens_per_sec": 106176.708213, "grad_norm": 33.0, "loss": 2.229048, "lr_stage": "cosine_decay", "mfu": 60.991993, "stage_name": "permutation_cycle", "status": "training", "step": 1245, "tokens_per_sec_per_gpu": 106176.708213, "vram_gb": 20.709735} {"adam_lr": 3.739653208784432e-05, "dt_ms": 155.371428, "global_tokens_per_sec": 105450.533987, "grad_norm": 30.5, "loss": 3.150123, "lr_stage": "cosine_decay", "mfu": 60.574851, "stage_name": "permutation_cycle", "status": "training", "step": 1246, "tokens_per_sec_per_gpu": 105450.533987, "vram_gb": 20.709735} {"adam_lr": 3.731210277075972e-05, "dt_ms": 153.861761, "global_tokens_per_sec": 106485.197385, "grad_norm": 32.75, "loss": 2.480419, "lr_stage": "cosine_decay", "mfu": 61.169202, "stage_name": "permutation_cycle", "status": "training", "step": 1247, "tokens_per_sec_per_gpu": 106485.197385, "vram_gb": 20.709735} {"adam_lr": 3.7227712103210486e-05, "dt_ms": 154.55699, "global_tokens_per_sec": 106006.205446, "grad_norm": 32.75, "loss": 2.083867, "lr_stage": "cosine_decay", "mfu": 60.89405, "stage_name": "permutation_cycle", "status": "training", "step": 1248, "tokens_per_sec_per_gpu": 106006.205446, "vram_gb": 20.709735} {"adam_lr": 3.71433603422652e-05, "dt_ms": 153.6057, "global_tokens_per_sec": 106662.7088, "grad_norm": 27.125, "loss": 1.87331, "lr_stage": "cosine_decay", "mfu": 61.271171, "stage_name": "permutation_cycle", "status": "training", "step": 1249, "tokens_per_sec_per_gpu": 106662.7088, "vram_gb": 20.709735} {"adam_lr": 3.705904774487396e-05, "dt_ms": 156.235933, "global_tokens_per_sec": 104867.04085, "grad_norm": 44.75, "loss": 3.428272, "lr_stage": "cosine_decay", "mfu": 60.23967, "stage_name": "permutation_cycle", "status": "training", "step": 1250, "tokens_per_sec_per_gpu": 104867.04085, "vram_gb": 20.709735} {"adam_lr": 3.6974774567867584e-05, "dt_ms": 155.896664, "global_tokens_per_sec": 105095.257427, "grad_norm": 27.625, "loss": 1.673117, "lr_stage": "cosine_decay", "mfu": 60.370766, "stage_name": "permutation_cycle", "status": "training", "step": 1251, "tokens_per_sec_per_gpu": 105095.257427, "vram_gb": 20.709735} {"adam_lr": 3.6890541067956776e-05, "dt_ms": 157.776117, "global_tokens_per_sec": 103843.346368, "grad_norm": 35.0, "loss": 2.323095, "lr_stage": "cosine_decay", "mfu": 59.651621, "stage_name": "permutation_cycle", "status": "training", "step": 1252, "tokens_per_sec_per_gpu": 103843.346368, "vram_gb": 20.709735} {"adam_lr": 3.680634750173137e-05, "dt_ms": 153.619289, "global_tokens_per_sec": 106653.272933, "grad_norm": 33.0, "loss": 2.388161, "lr_stage": "cosine_decay", "mfu": 61.265751, "stage_name": "permutation_cycle", "status": "training", "step": 1253, "tokens_per_sec_per_gpu": 106653.272933, "vram_gb": 20.709735} {"adam_lr": 3.6722194125659556e-05, "dt_ms": 156.826735, "global_tokens_per_sec": 104471.98335, "grad_norm": 20.5, "loss": 2.416481, "lr_stage": "cosine_decay", "mfu": 60.012734, "stage_name": "permutation_cycle", "status": "training", "step": 1254, "tokens_per_sec_per_gpu": 104471.98335, "vram_gb": 20.709735} {"adam_lr": 3.663808119608716e-05, "dt_ms": 156.200886, "global_tokens_per_sec": 104890.570364, "grad_norm": 24.375, "loss": 2.501816, "lr_stage": "cosine_decay", "mfu": 60.253186, "stage_name": "permutation_cycle", "status": "training", "step": 1255, "tokens_per_sec_per_gpu": 104890.570364, "vram_gb": 20.709735} {"adam_lr": 3.655400896923672e-05, "dt_ms": 155.076742, "global_tokens_per_sec": 105650.916898, "grad_norm": 35.0, "loss": 2.241845, "lr_stage": "cosine_decay", "mfu": 60.689959, "stage_name": "permutation_cycle", "status": "training", "step": 1256, "tokens_per_sec_per_gpu": 105650.916898, "vram_gb": 20.709735} {"adam_lr": 3.646997770120683e-05, "dt_ms": 153.194189, "global_tokens_per_sec": 106949.226333, "grad_norm": 37.75, "loss": 2.261253, "lr_stage": "cosine_decay", "mfu": 61.435758, "stage_name": "permutation_cycle", "status": "training", "step": 1257, "tokens_per_sec_per_gpu": 106949.226333, "vram_gb": 20.709735} {"adam_lr": 3.6385987647971285e-05, "dt_ms": 155.460119, "global_tokens_per_sec": 105390.373295, "grad_norm": 24.625, "loss": 2.199268, "lr_stage": "cosine_decay", "mfu": 60.540292, "stage_name": "permutation_cycle", "status": "training", "step": 1258, "tokens_per_sec_per_gpu": 105390.373295, "vram_gb": 20.709735} {"adam_lr": 3.630203906537838e-05, "dt_ms": 156.531572, "global_tokens_per_sec": 104668.979905, "grad_norm": 26.0, "loss": 2.337661, "lr_stage": "cosine_decay", "mfu": 60.125896, "stage_name": "permutation_cycle", "status": "training", "step": 1259, "tokens_per_sec_per_gpu": 104668.979905, "vram_gb": 20.709735} {"adam_lr": 3.6218132209150045e-05, "dt_ms": 156.948566, "global_tokens_per_sec": 104390.886594, "grad_norm": 31.375, "loss": 2.695698, "lr_stage": "cosine_decay", "mfu": 59.966149, "stage_name": "permutation_cycle", "status": "training", "step": 1260, "tokens_per_sec_per_gpu": 104390.886594, "vram_gb": 20.709735} {"adam_lr": 3.613426733488111e-05, "dt_ms": 156.39925, "global_tokens_per_sec": 104757.535582, "grad_norm": 29.125, "loss": 2.38719, "lr_stage": "cosine_decay", "mfu": 60.176766, "stage_name": "permutation_cycle", "status": "training", "step": 1261, "tokens_per_sec_per_gpu": 104757.535582, "vram_gb": 20.709735} {"adam_lr": 3.605044469803854e-05, "dt_ms": 155.435085, "global_tokens_per_sec": 105407.347181, "grad_norm": 33.0, "loss": 3.215186, "lr_stage": "cosine_decay", "mfu": 60.550043, "stage_name": "permutation_cycle", "status": "training", "step": 1262, "tokens_per_sec_per_gpu": 105407.347181, "vram_gb": 20.709735} {"adam_lr": 3.596666455396062e-05, "dt_ms": 154.588461, "global_tokens_per_sec": 105984.624611, "grad_norm": 29.375, "loss": 1.973849, "lr_stage": "cosine_decay", "mfu": 60.881653, "stage_name": "permutation_cycle", "status": "training", "step": 1263, "tokens_per_sec_per_gpu": 105984.624611, "vram_gb": 20.709735} {"adam_lr": 3.588292715785617e-05, "dt_ms": 154.553413, "global_tokens_per_sec": 106008.658369, "grad_norm": 28.625, "loss": 2.006505, "lr_stage": "cosine_decay", "mfu": 60.895459, "stage_name": "permutation_cycle", "status": "training", "step": 1264, "tokens_per_sec_per_gpu": 106008.658369, "vram_gb": 20.709735} {"adam_lr": 3.579923276480387e-05, "dt_ms": 155.070305, "global_tokens_per_sec": 105655.302694, "grad_norm": 20.625, "loss": 2.444599, "lr_stage": "cosine_decay", "mfu": 60.692478, "stage_name": "permutation_cycle", "status": "training", "step": 1265, "tokens_per_sec_per_gpu": 105655.302694, "vram_gb": 20.709735} {"adam_lr": 3.5715581629751326e-05, "dt_ms": 154.810905, "global_tokens_per_sec": 105832.337533, "grad_norm": 35.25, "loss": 2.161715, "lr_stage": "cosine_decay", "mfu": 60.794174, "stage_name": "permutation_cycle", "status": "training", "step": 1266, "tokens_per_sec_per_gpu": 105832.337533, "vram_gb": 20.709735} {"adam_lr": 3.563197400751441e-05, "dt_ms": 153.764963, "global_tokens_per_sec": 106552.231824, "grad_norm": 28.375, "loss": 2.303405, "lr_stage": "cosine_decay", "mfu": 61.207709, "stage_name": "permutation_cycle", "status": "training", "step": 1267, "tokens_per_sec_per_gpu": 106552.231824, "vram_gb": 20.709735} {"adam_lr": 3.554841015277641e-05, "dt_ms": 159.032822, "global_tokens_per_sec": 103022.758632, "grad_norm": 25.0, "loss": 2.907171, "lr_stage": "cosine_decay", "mfu": 59.180243, "stage_name": "permutation_cycle", "status": "training", "step": 1268, "tokens_per_sec_per_gpu": 103022.758632, "vram_gb": 20.709735} {"adam_lr": 3.5464890320087373e-05, "dt_ms": 158.918858, "global_tokens_per_sec": 103096.638436, "grad_norm": 26.375, "loss": 2.187426, "lr_stage": "cosine_decay", "mfu": 59.222683, "stage_name": "permutation_cycle", "status": "training", "step": 1269, "tokens_per_sec_per_gpu": 103096.638436, "vram_gb": 20.709735} {"adam_lr": 3.5381414763863166e-05, "dt_ms": 154.747486, "global_tokens_per_sec": 105875.710239, "grad_norm": 34.5, "loss": 1.78644, "lr_stage": "cosine_decay", "mfu": 60.819089, "stage_name": "permutation_cycle", "status": "training", "step": 1270, "tokens_per_sec_per_gpu": 105875.710239, "vram_gb": 20.709735} {"adam_lr": 3.529798373838481e-05, "dt_ms": 155.497789, "global_tokens_per_sec": 105364.841938, "grad_norm": 27.875, "loss": 2.73916, "lr_stage": "cosine_decay", "mfu": 60.525626, "stage_name": "permutation_cycle", "status": "training", "step": 1271, "tokens_per_sec_per_gpu": 105364.841938, "vram_gb": 20.709735} {"adam_lr": 3.5214597497797684e-05, "dt_ms": 155.771017, "global_tokens_per_sec": 105180.028401, "grad_norm": 24.0, "loss": 2.247027, "lr_stage": "cosine_decay", "mfu": 60.419462, "stage_name": "permutation_cycle", "status": "training", "step": 1272, "tokens_per_sec_per_gpu": 105180.028401, "vram_gb": 20.709735} {"adam_lr": 3.5131256296110704e-05, "dt_ms": 153.793573, "global_tokens_per_sec": 106532.409905, "grad_norm": 32.5, "loss": 1.987776, "lr_stage": "cosine_decay", "mfu": 61.196322, "stage_name": "permutation_cycle", "status": "training", "step": 1273, "tokens_per_sec_per_gpu": 106532.409905, "vram_gb": 20.709735} {"adam_lr": 3.504796038719567e-05, "dt_ms": 155.319691, "global_tokens_per_sec": 105485.659453, "grad_norm": 35.25, "loss": 3.426858, "lr_stage": "cosine_decay", "mfu": 60.595028, "stage_name": "permutation_cycle", "status": "training", "step": 1274, "tokens_per_sec_per_gpu": 105485.659453, "vram_gb": 20.709735} {"adam_lr": 3.4964710024786354e-05, "dt_ms": 154.256582, "global_tokens_per_sec": 106212.647525, "grad_norm": 25.625, "loss": 2.17097, "lr_stage": "cosine_decay", "mfu": 61.012638, "stage_name": "permutation_cycle", "status": "training", "step": 1275, "tokens_per_sec_per_gpu": 106212.647525, "vram_gb": 20.709735} {"adam_lr": 3.488150546247778e-05, "dt_ms": 153.427601, "global_tokens_per_sec": 106786.522817, "grad_norm": 30.875, "loss": 1.432027, "lr_stage": "cosine_decay", "mfu": 61.342295, "stage_name": "permutation_cycle", "status": "training", "step": 1276, "tokens_per_sec_per_gpu": 106786.522817, "vram_gb": 20.709735} {"adam_lr": 3.479834695372548e-05, "dt_ms": 155.005693, "global_tokens_per_sec": 105699.343275, "grad_norm": 26.5, "loss": 1.500785, "lr_stage": "cosine_decay", "mfu": 60.717777, "stage_name": "permutation_cycle", "status": "training", "step": 1277, "tokens_per_sec_per_gpu": 105699.343275, "vram_gb": 20.709735} {"adam_lr": 3.471523475184472e-05, "dt_ms": 155.348063, "global_tokens_per_sec": 105466.394204, "grad_norm": 22.375, "loss": 1.902357, "lr_stage": "cosine_decay", "mfu": 60.583962, "stage_name": "permutation_cycle", "status": "training", "step": 1278, "tokens_per_sec_per_gpu": 105466.394204, "vram_gb": 20.709735} {"adam_lr": 3.463216911000965e-05, "dt_ms": 155.35903, "global_tokens_per_sec": 105458.949018, "grad_norm": 16.375, "loss": 2.730475, "lr_stage": "cosine_decay", "mfu": 60.579685, "stage_name": "permutation_cycle", "status": "training", "step": 1279, "tokens_per_sec_per_gpu": 105458.949018, "vram_gb": 20.709735} {"adam_lr": 3.4549150281252636e-05, "dt_ms": 154.114485, "global_tokens_per_sec": 106310.578286, "grad_norm": 30.375, "loss": 1.368944, "lr_stage": "cosine_decay", "mfu": 61.068894, "stage_name": "permutation_cycle", "status": "training", "step": 1280, "tokens_per_sec_per_gpu": 106310.578286, "vram_gb": 20.709735} {"adam_lr": 3.4466178518463424e-05, "dt_ms": 153.497458, "global_tokens_per_sec": 106737.924304, "grad_norm": 38.0, "loss": 1.356988, "lr_stage": "cosine_decay", "mfu": 61.314378, "stage_name": "permutation_cycle", "status": "training", "step": 1281, "tokens_per_sec_per_gpu": 106737.924304, "vram_gb": 20.709735} {"adam_lr": 3.438325407438837e-05, "dt_ms": 156.95858, "global_tokens_per_sec": 104384.226706, "grad_norm": 28.0, "loss": 2.044684, "lr_stage": "cosine_decay", "mfu": 59.962323, "stage_name": "permutation_cycle", "status": "training", "step": 1282, "tokens_per_sec_per_gpu": 104384.226706, "vram_gb": 20.709735} {"adam_lr": 3.4300377201629754e-05, "dt_ms": 158.998966, "global_tokens_per_sec": 103044.695131, "grad_norm": 19.75, "loss": 2.505705, "lr_stage": "cosine_decay", "mfu": 59.192844, "stage_name": "permutation_cycle", "status": "training", "step": 1283, "tokens_per_sec_per_gpu": 103044.695131, "vram_gb": 20.709735} {"adam_lr": 3.4217548152644885e-05, "dt_ms": 155.068874, "global_tokens_per_sec": 105656.277365, "grad_norm": 33.25, "loss": 2.153073, "lr_stage": "cosine_decay", "mfu": 60.693038, "stage_name": "permutation_cycle", "status": "training", "step": 1284, "tokens_per_sec_per_gpu": 105656.277365, "vram_gb": 20.709735} {"adam_lr": 3.4134767179745406e-05, "dt_ms": 155.219555, "global_tokens_per_sec": 105553.710745, "grad_norm": 19.5, "loss": 1.75734, "lr_stage": "cosine_decay", "mfu": 60.63412, "stage_name": "permutation_cycle", "status": "training", "step": 1285, "tokens_per_sec_per_gpu": 105553.710745, "vram_gb": 20.709735} {"adam_lr": 3.40520345350965e-05, "dt_ms": 155.333996, "global_tokens_per_sec": 105475.945002, "grad_norm": 16.0, "loss": 1.998706, "lr_stage": "cosine_decay", "mfu": 60.589448, "stage_name": "permutation_cycle", "status": "training", "step": 1286, "tokens_per_sec_per_gpu": 105475.945002, "vram_gb": 20.709735} {"adam_lr": 3.396935047071619e-05, "dt_ms": 156.376123, "global_tokens_per_sec": 104773.028265, "grad_norm": 35.5, "loss": 2.983603, "lr_stage": "cosine_decay", "mfu": 60.185666, "stage_name": "permutation_cycle", "status": "training", "step": 1287, "tokens_per_sec_per_gpu": 104773.028265, "vram_gb": 20.709735} {"adam_lr": 3.388671523847445e-05, "dt_ms": 153.842449, "global_tokens_per_sec": 106498.564515, "grad_norm": 29.875, "loss": 2.728646, "lr_stage": "cosine_decay", "mfu": 61.17688, "stage_name": "permutation_cycle", "status": "training", "step": 1288, "tokens_per_sec_per_gpu": 106498.564515, "vram_gb": 20.709735} {"adam_lr": 3.380412909009254e-05, "dt_ms": 154.774666, "global_tokens_per_sec": 105857.117584, "grad_norm": 36.75, "loss": 2.592853, "lr_stage": "cosine_decay", "mfu": 60.808408, "stage_name": "permutation_cycle", "status": "training", "step": 1289, "tokens_per_sec_per_gpu": 105857.117584, "vram_gb": 20.709735} {"adam_lr": 3.372159227714218e-05, "dt_ms": 155.154943, "global_tokens_per_sec": 105597.666655, "grad_norm": 43.75, "loss": 2.697815, "lr_stage": "cosine_decay", "mfu": 60.65937, "stage_name": "permutation_cycle", "status": "training", "step": 1290, "tokens_per_sec_per_gpu": 105597.666655, "vram_gb": 20.709735} {"adam_lr": 3.363910505104481e-05, "dt_ms": 156.155586, "global_tokens_per_sec": 104920.998308, "grad_norm": 30.5, "loss": 2.432071, "lr_stage": "cosine_decay", "mfu": 60.270665, "stage_name": "permutation_cycle", "status": "training", "step": 1291, "tokens_per_sec_per_gpu": 104920.998308, "vram_gb": 20.709735} {"adam_lr": 3.355666766307084e-05, "dt_ms": 157.468319, "global_tokens_per_sec": 104046.325701, "grad_norm": 36.25, "loss": 3.451504, "lr_stage": "cosine_decay", "mfu": 59.76822, "stage_name": "permutation_cycle", "status": "training", "step": 1292, "tokens_per_sec_per_gpu": 104046.325701, "vram_gb": 20.709735} {"adam_lr": 3.347428036433885e-05, "dt_ms": 154.520273, "global_tokens_per_sec": 106031.394197, "grad_norm": 24.5, "loss": 1.769589, "lr_stage": "cosine_decay", "mfu": 60.90852, "stage_name": "permutation_cycle", "status": "training", "step": 1293, "tokens_per_sec_per_gpu": 106031.394197, "vram_gb": 20.709735} {"adam_lr": 3.339194340581485e-05, "dt_ms": 155.869961, "global_tokens_per_sec": 105113.261834, "grad_norm": 36.75, "loss": 3.020277, "lr_stage": "cosine_decay", "mfu": 60.381109, "stage_name": "permutation_cycle", "status": "training", "step": 1294, "tokens_per_sec_per_gpu": 105113.261834, "vram_gb": 20.709735} {"adam_lr": 3.330965703831146e-05, "dt_ms": 156.78525, "global_tokens_per_sec": 104499.626274, "grad_norm": 33.0, "loss": 2.450195, "lr_stage": "cosine_decay", "mfu": 60.028613, "stage_name": "permutation_cycle", "status": "training", "step": 1295, "tokens_per_sec_per_gpu": 104499.626274, "vram_gb": 20.709735} {"adam_lr": 3.322742151248725e-05, "dt_ms": 154.674292, "global_tokens_per_sec": 105925.812424, "grad_norm": 28.5, "loss": 1.960378, "lr_stage": "cosine_decay", "mfu": 60.847869, "stage_name": "permutation_cycle", "status": "training", "step": 1296, "tokens_per_sec_per_gpu": 105925.812424, "vram_gb": 20.709735} {"adam_lr": 3.31452370788459e-05, "dt_ms": 154.189825, "global_tokens_per_sec": 106258.632785, "grad_norm": 21.25, "loss": 1.659317, "lr_stage": "cosine_decay", "mfu": 61.039054, "stage_name": "permutation_cycle", "status": "training", "step": 1297, "tokens_per_sec_per_gpu": 106258.632785, "vram_gb": 20.709735} {"adam_lr": 3.3063103987735433e-05, "dt_ms": 156.985521, "global_tokens_per_sec": 104366.312655, "grad_norm": 25.25, "loss": 1.911399, "lr_stage": "cosine_decay", "mfu": 59.952033, "stage_name": "permutation_cycle", "status": "training", "step": 1298, "tokens_per_sec_per_gpu": 104366.312655, "vram_gb": 20.709735} {"adam_lr": 3.29810224893475e-05, "dt_ms": 156.481981, "global_tokens_per_sec": 104702.150792, "grad_norm": 35.5, "loss": 2.604627, "lr_stage": "cosine_decay", "mfu": 60.144951, "stage_name": "permutation_cycle", "status": "training", "step": 1299, "tokens_per_sec_per_gpu": 104702.150792, "vram_gb": 20.709735} {"adam_lr": 3.289899283371657e-05, "dt_ms": 156.120062, "global_tokens_per_sec": 104944.872576, "grad_norm": 36.75, "loss": 2.209944, "lr_stage": "cosine_decay", "mfu": 60.28438, "stage_name": "permutation_cycle", "status": "training", "step": 1300, "tokens_per_sec_per_gpu": 104944.872576, "vram_gb": 20.709735} {"adam_lr": 3.28170152707192e-05, "dt_ms": 158.138514, "global_tokens_per_sec": 103605.375001, "grad_norm": 27.125, "loss": 2.185721, "lr_stage": "cosine_decay", "mfu": 59.514921, "stage_name": "permutation_cycle", "status": "training", "step": 1301, "tokens_per_sec_per_gpu": 103605.375001, "vram_gb": 20.709735} {"adam_lr": 3.273509005007327e-05, "dt_ms": 156.120777, "global_tokens_per_sec": 104944.391779, "grad_norm": 31.375, "loss": 2.944697, "lr_stage": "cosine_decay", "mfu": 60.284103, "stage_name": "permutation_cycle", "status": "training", "step": 1302, "tokens_per_sec_per_gpu": 104944.391779, "vram_gb": 20.709735} {"adam_lr": 3.2653217421337215e-05, "dt_ms": 157.723665, "global_tokens_per_sec": 103877.880186, "grad_norm": 22.625, "loss": 2.332271, "lr_stage": "cosine_decay", "mfu": 59.671458, "stage_name": "permutation_cycle", "status": "training", "step": 1303, "tokens_per_sec_per_gpu": 103877.880186, "vram_gb": 20.709735} {"adam_lr": 3.257139763390925e-05, "dt_ms": 156.713963, "global_tokens_per_sec": 104547.161803, "grad_norm": 33.75, "loss": 2.988947, "lr_stage": "cosine_decay", "mfu": 60.055919, "stage_name": "permutation_cycle", "status": "training", "step": 1304, "tokens_per_sec_per_gpu": 104547.161803, "vram_gb": 20.709735} {"adam_lr": 3.248963093702663e-05, "dt_ms": 156.52585, "global_tokens_per_sec": 104672.806243, "grad_norm": 25.0, "loss": 2.283038, "lr_stage": "cosine_decay", "mfu": 60.128094, "stage_name": "permutation_cycle", "status": "training", "step": 1305, "tokens_per_sec_per_gpu": 104672.806243, "vram_gb": 20.709735} {"adam_lr": 3.240791757976491e-05, "dt_ms": 155.795336, "global_tokens_per_sec": 105163.610445, "grad_norm": 28.0, "loss": 2.773675, "lr_stage": "cosine_decay", "mfu": 60.410031, "stage_name": "permutation_cycle", "status": "training", "step": 1306, "tokens_per_sec_per_gpu": 105163.610445, "vram_gb": 20.709735} {"adam_lr": 3.2326257811037155e-05, "dt_ms": 155.404806, "global_tokens_per_sec": 105427.884808, "grad_norm": 33.5, "loss": 2.966085, "lr_stage": "cosine_decay", "mfu": 60.56184, "stage_name": "permutation_cycle", "status": "training", "step": 1307, "tokens_per_sec_per_gpu": 105427.884808, "vram_gb": 20.709735} {"adam_lr": 3.224465187959316e-05, "dt_ms": 154.300928, "global_tokens_per_sec": 106182.122169, "grad_norm": 28.875, "loss": 2.517826, "lr_stage": "cosine_decay", "mfu": 60.995103, "stage_name": "permutation_cycle", "status": "training", "step": 1308, "tokens_per_sec_per_gpu": 106182.122169, "vram_gb": 20.709735} {"adam_lr": 3.216310003401873e-05, "dt_ms": 155.065775, "global_tokens_per_sec": 105658.389214, "grad_norm": 31.0, "loss": 2.670727, "lr_stage": "cosine_decay", "mfu": 60.694251, "stage_name": "permutation_cycle", "status": "training", "step": 1309, "tokens_per_sec_per_gpu": 105658.389214, "vram_gb": 20.709735} {"adam_lr": 3.2081602522734986e-05, "dt_ms": 154.604435, "global_tokens_per_sec": 105973.674064, "grad_norm": 35.25, "loss": 1.93997, "lr_stage": "cosine_decay", "mfu": 60.875363, "stage_name": "permutation_cycle", "status": "training", "step": 1310, "tokens_per_sec_per_gpu": 105973.674064, "vram_gb": 20.709735} {"adam_lr": 3.200015959399745e-05, "dt_ms": 155.521154, "global_tokens_per_sec": 105349.012247, "grad_norm": 35.75, "loss": 2.378593, "lr_stage": "cosine_decay", "mfu": 60.516533, "stage_name": "permutation_cycle", "status": "training", "step": 1311, "tokens_per_sec_per_gpu": 105349.012247, "vram_gb": 20.709735} {"adam_lr": 3.1918771495895396e-05, "dt_ms": 156.646252, "global_tokens_per_sec": 104592.35267, "grad_norm": 27.0, "loss": 1.78056, "lr_stage": "cosine_decay", "mfu": 60.081879, "stage_name": "permutation_cycle", "status": "training", "step": 1312, "tokens_per_sec_per_gpu": 104592.35267, "vram_gb": 20.709735} {"adam_lr": 3.183743847635109e-05, "dt_ms": 156.422853, "global_tokens_per_sec": 104741.728185, "grad_norm": 30.5, "loss": 2.094382, "lr_stage": "cosine_decay", "mfu": 60.167686, "stage_name": "permutation_cycle", "status": "training", "step": 1313, "tokens_per_sec_per_gpu": 104741.728185, "vram_gb": 20.709735} {"adam_lr": 3.1756160783119016e-05, "dt_ms": 157.37772, "global_tokens_per_sec": 104106.222994, "grad_norm": 25.375, "loss": 1.663322, "lr_stage": "cosine_decay", "mfu": 59.802627, "stage_name": "permutation_cycle", "status": "training", "step": 1314, "tokens_per_sec_per_gpu": 104106.222994, "vram_gb": 20.709735} {"adam_lr": 3.167493866378514e-05, "dt_ms": 156.006336, "global_tokens_per_sec": 105021.375399, "grad_norm": 35.25, "loss": 2.55966, "lr_stage": "cosine_decay", "mfu": 60.328326, "stage_name": "permutation_cycle", "status": "training", "step": 1315, "tokens_per_sec_per_gpu": 105021.375399, "vram_gb": 20.709735} {"adam_lr": 3.1593772365766105e-05, "dt_ms": 157.299519, "global_tokens_per_sec": 104157.979296, "grad_norm": 38.5, "loss": 3.655775, "lr_stage": "cosine_decay", "mfu": 59.832358, "stage_name": "permutation_cycle", "status": "training", "step": 1316, "tokens_per_sec_per_gpu": 104157.979296, "vram_gb": 20.709735} {"adam_lr": 3.151266213630854e-05, "dt_ms": 155.004501, "global_tokens_per_sec": 105700.156177, "grad_norm": 38.0, "loss": 2.934536, "lr_stage": "cosine_decay", "mfu": 60.718244, "stage_name": "permutation_cycle", "status": "training", "step": 1317, "tokens_per_sec_per_gpu": 105700.156177, "vram_gb": 20.709735} {"adam_lr": 3.1431608222488275e-05, "dt_ms": 153.623104, "global_tokens_per_sec": 106650.624569, "grad_norm": 32.75, "loss": 2.492047, "lr_stage": "cosine_decay", "mfu": 61.264229, "stage_name": "permutation_cycle", "status": "training", "step": 1318, "tokens_per_sec_per_gpu": 106650.624569, "vram_gb": 20.709735} {"adam_lr": 3.135061087120955e-05, "dt_ms": 154.926777, "global_tokens_per_sec": 105753.184371, "grad_norm": 25.0, "loss": 1.846698, "lr_stage": "cosine_decay", "mfu": 60.748705, "stage_name": "permutation_cycle", "status": "training", "step": 1319, "tokens_per_sec_per_gpu": 105753.184371, "vram_gb": 20.709735} {"adam_lr": 3.12696703292044e-05, "dt_ms": 156.071663, "global_tokens_per_sec": 104977.416754, "grad_norm": 29.375, "loss": 2.607602, "lr_stage": "cosine_decay", "mfu": 60.303074, "stage_name": "permutation_cycle", "status": "training", "step": 1320, "tokens_per_sec_per_gpu": 104977.416754, "vram_gb": 20.709735} {"adam_lr": 3.118878684303172e-05, "dt_ms": 155.282021, "global_tokens_per_sec": 105511.249403, "grad_norm": 30.5, "loss": 2.880288, "lr_stage": "cosine_decay", "mfu": 60.609728, "stage_name": "permutation_cycle", "status": "training", "step": 1321, "tokens_per_sec_per_gpu": 105511.249403, "vram_gb": 20.709735} {"adam_lr": 3.110796065907665e-05, "dt_ms": 156.271219, "global_tokens_per_sec": 104843.361934, "grad_norm": 26.375, "loss": 2.145241, "lr_stage": "cosine_decay", "mfu": 60.226068, "stage_name": "permutation_cycle", "status": "training", "step": 1322, "tokens_per_sec_per_gpu": 104843.361934, "vram_gb": 20.709735} {"adam_lr": 3.102719202354974e-05, "dt_ms": 154.639959, "global_tokens_per_sec": 105949.329465, "grad_norm": 28.5, "loss": 2.256731, "lr_stage": "cosine_decay", "mfu": 60.861378, "stage_name": "permutation_cycle", "status": "training", "step": 1323, "tokens_per_sec_per_gpu": 105949.329465, "vram_gb": 20.709735} {"adam_lr": 3.09464811824863e-05, "dt_ms": 156.912804, "global_tokens_per_sec": 104414.678846, "grad_norm": 31.0, "loss": 2.944461, "lr_stage": "cosine_decay", "mfu": 59.979816, "stage_name": "permutation_cycle", "status": "training", "step": 1324, "tokens_per_sec_per_gpu": 104414.678846, "vram_gb": 20.709735} {"adam_lr": 3.086582838174551e-05, "dt_ms": 157.410622, "global_tokens_per_sec": 104084.462846, "grad_norm": 38.75, "loss": 2.928188, "lr_stage": "cosine_decay", "mfu": 59.790127, "stage_name": "permutation_cycle", "status": "training", "step": 1325, "tokens_per_sec_per_gpu": 104084.462846, "vram_gb": 20.709735} {"adam_lr": 3.078523386700982e-05, "dt_ms": 154.910564, "global_tokens_per_sec": 105764.252174, "grad_norm": 20.125, "loss": 2.216038, "lr_stage": "cosine_decay", "mfu": 60.755063, "stage_name": "permutation_cycle", "status": "training", "step": 1326, "tokens_per_sec_per_gpu": 105764.252174, "vram_gb": 20.709735} {"adam_lr": 3.070469788378408e-05, "dt_ms": 154.778242, "global_tokens_per_sec": 105854.671668, "grad_norm": 37.0, "loss": 2.281712, "lr_stage": "cosine_decay", "mfu": 60.807003, "stage_name": "permutation_cycle", "status": "training", "step": 1327, "tokens_per_sec_per_gpu": 105854.671668, "vram_gb": 20.709735} {"adam_lr": 3.062422067739485e-05, "dt_ms": 156.783819, "global_tokens_per_sec": 104500.579739, "grad_norm": 30.75, "loss": 2.38045, "lr_stage": "cosine_decay", "mfu": 60.029161, "stage_name": "permutation_cycle", "status": "training", "step": 1328, "tokens_per_sec_per_gpu": 104500.579739, "vram_gb": 20.709735} {"adam_lr": 3.054380249298969e-05, "dt_ms": 155.283213, "global_tokens_per_sec": 105510.439404, "grad_norm": 34.0, "loss": 2.356909, "lr_stage": "cosine_decay", "mfu": 60.609263, "stage_name": "permutation_cycle", "status": "training", "step": 1329, "tokens_per_sec_per_gpu": 105510.439404, "vram_gb": 20.709735} {"adam_lr": 3.046344357553632e-05, "dt_ms": 154.623508, "global_tokens_per_sec": 105960.601747, "grad_norm": 23.25, "loss": 2.154895, "lr_stage": "cosine_decay", "mfu": 60.867854, "stage_name": "permutation_cycle", "status": "training", "step": 1330, "tokens_per_sec_per_gpu": 105960.601747, "vram_gb": 20.709735} {"adam_lr": 3.038314416982194e-05, "dt_ms": 154.188156, "global_tokens_per_sec": 106259.782927, "grad_norm": 49.0, "loss": 2.42805, "lr_stage": "cosine_decay", "mfu": 61.039715, "stage_name": "permutation_cycle", "status": "training", "step": 1331, "tokens_per_sec_per_gpu": 106259.782927, "vram_gb": 20.709735} {"adam_lr": 3.0302904520452447e-05, "dt_ms": 155.240297, "global_tokens_per_sec": 105539.607197, "grad_norm": 42.25, "loss": 2.861164, "lr_stage": "cosine_decay", "mfu": 60.626018, "stage_name": "permutation_cycle", "status": "training", "step": 1332, "tokens_per_sec_per_gpu": 105539.607197, "vram_gb": 20.709735} {"adam_lr": 3.0222724871851758e-05, "dt_ms": 154.723883, "global_tokens_per_sec": 105891.861791, "grad_norm": 29.75, "loss": 1.944401, "lr_stage": "cosine_decay", "mfu": 60.828367, "stage_name": "permutation_cycle", "status": "training", "step": 1333, "tokens_per_sec_per_gpu": 105891.861791, "vram_gb": 20.709735} {"adam_lr": 3.0142605468260978e-05, "dt_ms": 154.836655, "global_tokens_per_sec": 105814.737703, "grad_norm": 29.75, "loss": 2.692969, "lr_stage": "cosine_decay", "mfu": 60.784064, "stage_name": "permutation_cycle", "status": "training", "step": 1334, "tokens_per_sec_per_gpu": 105814.737703, "vram_gb": 20.709735} {"adam_lr": 3.006254655373769e-05, "dt_ms": 155.713558, "global_tokens_per_sec": 105218.84022, "grad_norm": 34.0, "loss": 3.179887, "lr_stage": "cosine_decay", "mfu": 60.441757, "stage_name": "permutation_cycle", "status": "training", "step": 1335, "tokens_per_sec_per_gpu": 105218.84022, "vram_gb": 20.709735} {"adam_lr": 2.9982548372155263e-05, "dt_ms": 155.734301, "global_tokens_per_sec": 105204.826011, "grad_norm": 37.0, "loss": 2.350719, "lr_stage": "cosine_decay", "mfu": 60.433707, "stage_name": "permutation_cycle", "status": "training", "step": 1336, "tokens_per_sec_per_gpu": 105204.826011, "vram_gb": 20.709735} {"adam_lr": 2.9902611167202e-05, "dt_ms": 154.519558, "global_tokens_per_sec": 106031.885006, "grad_norm": 19.125, "loss": 1.687529, "lr_stage": "cosine_decay", "mfu": 60.908801, "stage_name": "permutation_cycle", "status": "training", "step": 1337, "tokens_per_sec_per_gpu": 106031.885006, "vram_gb": 20.709735} {"adam_lr": 2.9822735182380496e-05, "dt_ms": 155.687809, "global_tokens_per_sec": 105236.242364, "grad_norm": 25.5, "loss": 3.339623, "lr_stage": "cosine_decay", "mfu": 60.451754, "stage_name": "permutation_cycle", "status": "training", "step": 1338, "tokens_per_sec_per_gpu": 105236.242364, "vram_gb": 20.709735} {"adam_lr": 2.9742920661006878e-05, "dt_ms": 158.045292, "global_tokens_per_sec": 103666.485746, "grad_norm": 28.5, "loss": 2.354726, "lr_stage": "cosine_decay", "mfu": 59.550025, "stage_name": "permutation_cycle", "status": "training", "step": 1339, "tokens_per_sec_per_gpu": 103666.485746, "vram_gb": 20.709735} {"adam_lr": 2.9663167846209998e-05, "dt_ms": 157.730103, "global_tokens_per_sec": 103873.640708, "grad_norm": 33.0, "loss": 2.704108, "lr_stage": "cosine_decay", "mfu": 59.669023, "stage_name": "permutation_cycle", "status": "training", "step": 1340, "tokens_per_sec_per_gpu": 103873.640708, "vram_gb": 20.709735} {"adam_lr": 2.958347698093077e-05, "dt_ms": 157.325745, "global_tokens_per_sec": 104140.616265, "grad_norm": 31.625, "loss": 2.447268, "lr_stage": "cosine_decay", "mfu": 59.822384, "stage_name": "permutation_cycle", "status": "training", "step": 1341, "tokens_per_sec_per_gpu": 104140.616265, "vram_gb": 20.709735} {"adam_lr": 2.950384830792136e-05, "dt_ms": 154.570103, "global_tokens_per_sec": 105997.212363, "grad_norm": 47.75, "loss": 3.534307, "lr_stage": "cosine_decay", "mfu": 60.888884, "stage_name": "permutation_cycle", "status": "training", "step": 1342, "tokens_per_sec_per_gpu": 105997.212363, "vram_gb": 20.709735} {"adam_lr": 2.9424282069744564e-05, "dt_ms": 156.09026, "global_tokens_per_sec": 104964.909707, "grad_norm": 26.75, "loss": 2.968815, "lr_stage": "cosine_decay", "mfu": 60.29589, "stage_name": "permutation_cycle", "status": "training", "step": 1343, "tokens_per_sec_per_gpu": 104964.909707, "vram_gb": 20.709735} {"adam_lr": 2.934477850877292e-05, "dt_ms": 155.405045, "global_tokens_per_sec": 105427.723063, "grad_norm": 28.375, "loss": 2.517379, "lr_stage": "cosine_decay", "mfu": 60.561748, "stage_name": "permutation_cycle", "status": "training", "step": 1344, "tokens_per_sec_per_gpu": 105427.723063, "vram_gb": 20.709735} {"adam_lr": 2.926533786718806e-05, "dt_ms": 154.230356, "global_tokens_per_sec": 106230.708415, "grad_norm": 35.25, "loss": 2.695635, "lr_stage": "cosine_decay", "mfu": 61.023013, "stage_name": "permutation_cycle", "status": "training", "step": 1345, "tokens_per_sec_per_gpu": 106230.708415, "vram_gb": 20.709735} {"adam_lr": 2.918596038697995e-05, "dt_ms": 155.677319, "global_tokens_per_sec": 105243.333776, "grad_norm": 25.25, "loss": 2.198186, "lr_stage": "cosine_decay", "mfu": 60.455827, "stage_name": "permutation_cycle", "status": "training", "step": 1346, "tokens_per_sec_per_gpu": 105243.333776, "vram_gb": 20.709735} {"adam_lr": 2.910664630994615e-05, "dt_ms": 153.928041, "global_tokens_per_sec": 106439.345585, "grad_norm": 37.0, "loss": 1.986773, "lr_stage": "cosine_decay", "mfu": 61.142863, "stage_name": "permutation_cycle", "status": "training", "step": 1347, "tokens_per_sec_per_gpu": 106439.345585, "vram_gb": 20.709735} {"adam_lr": 2.9027395877691144e-05, "dt_ms": 156.685352, "global_tokens_per_sec": 104566.25177, "grad_norm": 26.5, "loss": 2.884454, "lr_stage": "cosine_decay", "mfu": 60.066885, "stage_name": "permutation_cycle", "status": "training", "step": 1348, "tokens_per_sec_per_gpu": 104566.25177, "vram_gb": 20.709735} {"adam_lr": 2.8948209331625454e-05, "dt_ms": 155.450583, "global_tokens_per_sec": 105396.838893, "grad_norm": 45.25, "loss": 2.175551, "lr_stage": "cosine_decay", "mfu": 60.544006, "stage_name": "permutation_cycle", "status": "training", "step": 1349, "tokens_per_sec_per_gpu": 105396.838893, "vram_gb": 20.709735} {"adam_lr": 2.886908691296504e-05, "dt_ms": 153.773546, "global_tokens_per_sec": 106546.284474, "grad_norm": 15.875, "loss": 1.701839, "lr_stage": "cosine_decay", "mfu": 61.204292, "stage_name": "permutation_cycle", "status": "training", "step": 1350, "tokens_per_sec_per_gpu": 106546.284474, "vram_gb": 20.709735} {"adam_lr": 2.8790028862730484e-05, "dt_ms": 155.460835, "global_tokens_per_sec": 105389.888407, "grad_norm": 33.75, "loss": 2.170016, "lr_stage": "cosine_decay", "mfu": 60.540014, "stage_name": "permutation_cycle", "status": "training", "step": 1351, "tokens_per_sec_per_gpu": 105389.888407, "vram_gb": 20.709735} {"adam_lr": 2.8711035421746367e-05, "dt_ms": 154.399872, "global_tokens_per_sec": 106114.077727, "grad_norm": 15.4375, "loss": 1.949749, "lr_stage": "cosine_decay", "mfu": 60.956016, "stage_name": "permutation_cycle", "status": "training", "step": 1352, "tokens_per_sec_per_gpu": 106114.077727, "vram_gb": 20.709735} {"adam_lr": 2.8632106830640383e-05, "dt_ms": 154.628515, "global_tokens_per_sec": 105957.170799, "grad_norm": 40.25, "loss": 2.516073, "lr_stage": "cosine_decay", "mfu": 60.865883, "stage_name": "permutation_cycle", "status": "training", "step": 1353, "tokens_per_sec_per_gpu": 105957.170799, "vram_gb": 20.709735} {"adam_lr": 2.8553243329842714e-05, "dt_ms": 154.307127, "global_tokens_per_sec": 106177.856582, "grad_norm": 34.5, "loss": 2.089702, "lr_stage": "cosine_decay", "mfu": 60.992653, "stage_name": "permutation_cycle", "status": "training", "step": 1354, "tokens_per_sec_per_gpu": 106177.856582, "vram_gb": 20.709735} {"adam_lr": 2.8474445159585235e-05, "dt_ms": 158.214569, "global_tokens_per_sec": 103555.570729, "grad_norm": 45.25, "loss": 3.745538, "lr_stage": "cosine_decay", "mfu": 59.486311, "stage_name": "permutation_cycle", "status": "training", "step": 1355, "tokens_per_sec_per_gpu": 103555.570729, "vram_gb": 20.709735} {"adam_lr": 2.8395712559900877e-05, "dt_ms": 158.871889, "global_tokens_per_sec": 103127.11765, "grad_norm": 27.375, "loss": 2.66641, "lr_stage": "cosine_decay", "mfu": 59.240191, "stage_name": "permutation_cycle", "status": "training", "step": 1356, "tokens_per_sec_per_gpu": 103127.11765, "vram_gb": 20.709735} {"adam_lr": 2.8317045770622786e-05, "dt_ms": 153.742313, "global_tokens_per_sec": 106567.929409, "grad_norm": 34.25, "loss": 2.111834, "lr_stage": "cosine_decay", "mfu": 61.216726, "stage_name": "permutation_cycle", "status": "training", "step": 1357, "tokens_per_sec_per_gpu": 106567.929409, "vram_gb": 20.709735} {"adam_lr": 2.823844503138363e-05, "dt_ms": 156.352043, "global_tokens_per_sec": 104789.164693, "grad_norm": 20.625, "loss": 2.243166, "lr_stage": "cosine_decay", "mfu": 60.194935, "stage_name": "permutation_cycle", "status": "training", "step": 1358, "tokens_per_sec_per_gpu": 104789.164693, "vram_gb": 20.709735} {"adam_lr": 2.81599105816149e-05, "dt_ms": 154.984713, "global_tokens_per_sec": 105713.652173, "grad_norm": 21.75, "loss": 2.319746, "lr_stage": "cosine_decay", "mfu": 60.725996, "stage_name": "permutation_cycle", "status": "training", "step": 1359, "tokens_per_sec_per_gpu": 105713.652173, "vram_gb": 20.709735} {"adam_lr": 2.8081442660546125e-05, "dt_ms": 154.582977, "global_tokens_per_sec": 105988.384276, "grad_norm": 29.375, "loss": 2.587012, "lr_stage": "cosine_decay", "mfu": 60.883813, "stage_name": "permutation_cycle", "status": "training", "step": 1360, "tokens_per_sec_per_gpu": 105988.384276, "vram_gb": 20.709735} {"adam_lr": 2.8003041507204242e-05, "dt_ms": 155.479431, "global_tokens_per_sec": 105377.282889, "grad_norm": 36.75, "loss": 3.081051, "lr_stage": "cosine_decay", "mfu": 60.532773, "stage_name": "permutation_cycle", "status": "training", "step": 1361, "tokens_per_sec_per_gpu": 105377.282889, "vram_gb": 20.709735} {"adam_lr": 2.7924707360412746e-05, "dt_ms": 155.947208, "global_tokens_per_sec": 105061.194539, "grad_norm": 28.625, "loss": 2.846364, "lr_stage": "cosine_decay", "mfu": 60.351199, "stage_name": "permutation_cycle", "status": "training", "step": 1362, "tokens_per_sec_per_gpu": 105061.194539, "vram_gb": 20.709735} {"adam_lr": 2.7846440458791023e-05, "dt_ms": 155.982256, "global_tokens_per_sec": 105037.588421, "grad_norm": 24.375, "loss": 2.624923, "lr_stage": "cosine_decay", "mfu": 60.337639, "stage_name": "permutation_cycle", "status": "training", "step": 1363, "tokens_per_sec_per_gpu": 105037.588421, "vram_gb": 20.709735} {"adam_lr": 2.776824104075364e-05, "dt_ms": 154.039383, "global_tokens_per_sec": 106362.409975, "grad_norm": 29.25, "loss": 2.141392, "lr_stage": "cosine_decay", "mfu": 61.098668, "stage_name": "permutation_cycle", "status": "training", "step": 1364, "tokens_per_sec_per_gpu": 106362.409975, "vram_gb": 20.709735} {"adam_lr": 2.7690109344509563e-05, "dt_ms": 154.368639, "global_tokens_per_sec": 106135.547395, "grad_norm": 26.0, "loss": 1.982361, "lr_stage": "cosine_decay", "mfu": 60.968349, "stage_name": "permutation_cycle", "status": "training", "step": 1365, "tokens_per_sec_per_gpu": 106135.547395, "vram_gb": 20.709735} {"adam_lr": 2.761204560806152e-05, "dt_ms": 155.256748, "global_tokens_per_sec": 105528.424304, "grad_norm": 35.0, "loss": 2.694283, "lr_stage": "cosine_decay", "mfu": 60.619594, "stage_name": "permutation_cycle", "status": "training", "step": 1366, "tokens_per_sec_per_gpu": 105528.424304, "vram_gb": 20.709735} {"adam_lr": 2.753405006920518e-05, "dt_ms": 154.805183, "global_tokens_per_sec": 105836.249401, "grad_norm": 27.75, "loss": 1.633382, "lr_stage": "cosine_decay", "mfu": 60.796421, "stage_name": "permutation_cycle", "status": "training", "step": 1367, "tokens_per_sec_per_gpu": 105836.249401, "vram_gb": 20.709735} {"adam_lr": 2.7456122965528475e-05, "dt_ms": 155.700207, "global_tokens_per_sec": 105227.862835, "grad_norm": 33.0, "loss": 2.14455, "lr_stage": "cosine_decay", "mfu": 60.44694, "stage_name": "permutation_cycle", "status": "training", "step": 1368, "tokens_per_sec_per_gpu": 105227.862835, "vram_gb": 20.709735} {"adam_lr": 2.7378264534410864e-05, "dt_ms": 154.027224, "global_tokens_per_sec": 106370.806527, "grad_norm": 38.75, "loss": 2.752533, "lr_stage": "cosine_decay", "mfu": 61.103491, "stage_name": "permutation_cycle", "status": "training", "step": 1369, "tokens_per_sec_per_gpu": 106370.806527, "vram_gb": 20.709735} {"adam_lr": 2.7300475013022663e-05, "dt_ms": 155.383348, "global_tokens_per_sec": 105442.443877, "grad_norm": 31.0, "loss": 2.16624, "lr_stage": "cosine_decay", "mfu": 60.570204, "stage_name": "permutation_cycle", "status": "training", "step": 1370, "tokens_per_sec_per_gpu": 105442.443877, "vram_gb": 20.709735} {"adam_lr": 2.7222754638324233e-05, "dt_ms": 155.287027, "global_tokens_per_sec": 105507.847491, "grad_norm": 27.75, "loss": 2.189167, "lr_stage": "cosine_decay", "mfu": 60.607774, "stage_name": "permutation_cycle", "status": "training", "step": 1371, "tokens_per_sec_per_gpu": 105507.847491, "vram_gb": 20.709735} {"adam_lr": 2.7145103647065308e-05, "dt_ms": 155.700922, "global_tokens_per_sec": 105227.379442, "grad_norm": 36.0, "loss": 2.981521, "lr_stage": "cosine_decay", "mfu": 60.446662, "stage_name": "permutation_cycle", "status": "training", "step": 1372, "tokens_per_sec_per_gpu": 105227.379442, "vram_gb": 20.709735} {"adam_lr": 2.7067522275784273e-05, "dt_ms": 155.78866, "global_tokens_per_sec": 105168.116824, "grad_norm": 32.5, "loss": 2.086259, "lr_stage": "cosine_decay", "mfu": 60.41262, "stage_name": "permutation_cycle", "status": "training", "step": 1373, "tokens_per_sec_per_gpu": 105168.116824, "vram_gb": 20.709735} {"adam_lr": 2.699001076080742e-05, "dt_ms": 156.463861, "global_tokens_per_sec": 104714.276169, "grad_norm": 19.5, "loss": 2.445261, "lr_stage": "cosine_decay", "mfu": 60.151916, "stage_name": "permutation_cycle", "status": "training", "step": 1374, "tokens_per_sec_per_gpu": 104714.276169, "vram_gb": 20.709735} {"adam_lr": 2.6912569338248315e-05, "dt_ms": 154.219866, "global_tokens_per_sec": 106237.934491, "grad_norm": 25.375, "loss": 2.225334, "lr_stage": "cosine_decay", "mfu": 61.027164, "stage_name": "permutation_cycle", "status": "training", "step": 1375, "tokens_per_sec_per_gpu": 106237.934491, "vram_gb": 20.709735} {"adam_lr": 2.6835198244006927e-05, "dt_ms": 156.479359, "global_tokens_per_sec": 104703.905607, "grad_norm": 27.875, "loss": 2.614861, "lr_stage": "cosine_decay", "mfu": 60.145959, "stage_name": "permutation_cycle", "status": "training", "step": 1376, "tokens_per_sec_per_gpu": 104703.905607, "vram_gb": 20.709735} {"adam_lr": 2.6757897713769035e-05, "dt_ms": 155.201912, "global_tokens_per_sec": 105565.709833, "grad_norm": 21.0, "loss": 2.876844, "lr_stage": "cosine_decay", "mfu": 60.641012, "stage_name": "permutation_cycle", "status": "training", "step": 1377, "tokens_per_sec_per_gpu": 105565.709833, "vram_gb": 20.709735} {"adam_lr": 2.668066798300545e-05, "dt_ms": 155.198097, "global_tokens_per_sec": 105568.30459, "grad_norm": 25.0, "loss": 2.106339, "lr_stage": "cosine_decay", "mfu": 60.642503, "stage_name": "permutation_cycle", "status": "training", "step": 1378, "tokens_per_sec_per_gpu": 105568.30459, "vram_gb": 20.709735} {"adam_lr": 2.660350928697134e-05, "dt_ms": 154.533386, "global_tokens_per_sec": 106022.39684, "grad_norm": 28.5, "loss": 2.374049, "lr_stage": "cosine_decay", "mfu": 60.903351, "stage_name": "permutation_cycle", "status": "training", "step": 1379, "tokens_per_sec_per_gpu": 106022.39684, "vram_gb": 20.709735} {"adam_lr": 2.6526421860705473e-05, "dt_ms": 155.043364, "global_tokens_per_sec": 105673.662017, "grad_norm": 43.5, "loss": 3.049109, "lr_stage": "cosine_decay", "mfu": 60.703024, "stage_name": "permutation_cycle", "status": "training", "step": 1380, "tokens_per_sec_per_gpu": 105673.662017, "vram_gb": 20.709735} {"adam_lr": 2.64494059390295e-05, "dt_ms": 154.906988, "global_tokens_per_sec": 105766.693913, "grad_norm": 35.5, "loss": 2.933418, "lr_stage": "cosine_decay", "mfu": 60.756465, "stage_name": "permutation_cycle", "status": "training", "step": 1381, "tokens_per_sec_per_gpu": 105766.693913, "vram_gb": 20.709735} {"adam_lr": 2.6372461756547306e-05, "dt_ms": 155.15995, "global_tokens_per_sec": 105594.259169, "grad_norm": 29.75, "loss": 1.915323, "lr_stage": "cosine_decay", "mfu": 60.657412, "stage_name": "permutation_cycle", "status": "training", "step": 1382, "tokens_per_sec_per_gpu": 105594.259169, "vram_gb": 20.709735} {"adam_lr": 2.6295589547644195e-05, "dt_ms": 153.730869, "global_tokens_per_sec": 106575.862579, "grad_norm": 26.75, "loss": 2.029283, "lr_stage": "cosine_decay", "mfu": 61.221283, "stage_name": "permutation_cycle", "status": "training", "step": 1383, "tokens_per_sec_per_gpu": 106575.862579, "vram_gb": 20.709735} {"adam_lr": 2.6218789546486234e-05, "dt_ms": 153.7714, "global_tokens_per_sec": 106547.771249, "grad_norm": 35.5, "loss": 1.681947, "lr_stage": "cosine_decay", "mfu": 61.205146, "stage_name": "permutation_cycle", "status": "training", "step": 1384, "tokens_per_sec_per_gpu": 106547.771249, "vram_gb": 20.709735} {"adam_lr": 2.6142061987019577e-05, "dt_ms": 154.563189, "global_tokens_per_sec": 106001.95398, "grad_norm": 13.0625, "loss": 1.859027, "lr_stage": "cosine_decay", "mfu": 60.891608, "stage_name": "permutation_cycle", "status": "training", "step": 1385, "tokens_per_sec_per_gpu": 106001.95398, "vram_gb": 20.709735} {"adam_lr": 2.6065407102969664e-05, "dt_ms": 155.321836, "global_tokens_per_sec": 105484.202171, "grad_norm": 37.5, "loss": 3.704248, "lr_stage": "cosine_decay", "mfu": 60.594191, "stage_name": "permutation_cycle", "status": "training", "step": 1386, "tokens_per_sec_per_gpu": 105484.202171, "vram_gb": 20.709735} {"adam_lr": 2.5988825127840544e-05, "dt_ms": 155.340433, "global_tokens_per_sec": 105471.574083, "grad_norm": 21.5, "loss": 2.293523, "lr_stage": "cosine_decay", "mfu": 60.586937, "stage_name": "permutation_cycle", "status": "training", "step": 1387, "tokens_per_sec_per_gpu": 105471.574083, "vram_gb": 20.709735} {"adam_lr": 2.591231629491423e-05, "dt_ms": 155.883312, "global_tokens_per_sec": 105104.258859, "grad_norm": 25.75, "loss": 2.453842, "lr_stage": "cosine_decay", "mfu": 60.375937, "stage_name": "permutation_cycle", "status": "training", "step": 1388, "tokens_per_sec_per_gpu": 105104.258859, "vram_gb": 20.709735} {"adam_lr": 2.5835880837249882e-05, "dt_ms": 154.825449, "global_tokens_per_sec": 105822.396169, "grad_norm": 26.125, "loss": 1.707057, "lr_stage": "cosine_decay", "mfu": 60.788463, "stage_name": "permutation_cycle", "status": "training", "step": 1389, "tokens_per_sec_per_gpu": 105822.396169, "vram_gb": 20.709735} {"adam_lr": 2.575951898768315e-05, "dt_ms": 157.201767, "global_tokens_per_sec": 104222.747085, "grad_norm": 29.125, "loss": 2.829671, "lr_stage": "cosine_decay", "mfu": 59.869563, "stage_name": "permutation_cycle", "status": "training", "step": 1390, "tokens_per_sec_per_gpu": 104222.747085, "vram_gb": 20.709735} {"adam_lr": 2.5683230978825477e-05, "dt_ms": 156.17156, "global_tokens_per_sec": 104910.266439, "grad_norm": 15.4375, "loss": 1.997077, "lr_stage": "cosine_decay", "mfu": 60.264501, "stage_name": "permutation_cycle", "status": "training", "step": 1391, "tokens_per_sec_per_gpu": 104910.266439, "vram_gb": 20.709735} {"adam_lr": 2.560701704306336e-05, "dt_ms": 155.330896, "global_tokens_per_sec": 105478.049648, "grad_norm": 35.25, "loss": 2.16152, "lr_stage": "cosine_decay", "mfu": 60.590657, "stage_name": "permutation_cycle", "status": "training", "step": 1392, "tokens_per_sec_per_gpu": 105478.049648, "vram_gb": 20.709735} {"adam_lr": 2.5530877412557687e-05, "dt_ms": 155.044794, "global_tokens_per_sec": 105672.687025, "grad_norm": 30.75, "loss": 2.114768, "lr_stage": "cosine_decay", "mfu": 60.702464, "stage_name": "permutation_cycle", "status": "training", "step": 1393, "tokens_per_sec_per_gpu": 105672.687025, "vram_gb": 20.709735} {"adam_lr": 2.545481231924296e-05, "dt_ms": 155.388117, "global_tokens_per_sec": 105439.208181, "grad_norm": 26.875, "loss": 1.944243, "lr_stage": "cosine_decay", "mfu": 60.568345, "stage_name": "permutation_cycle", "status": "training", "step": 1394, "tokens_per_sec_per_gpu": 105439.208181, "vram_gb": 20.709735} {"adam_lr": 2.537882199482665e-05, "dt_ms": 155.007601, "global_tokens_per_sec": 105698.042658, "grad_norm": 28.0, "loss": 2.234934, "lr_stage": "cosine_decay", "mfu": 60.71703, "stage_name": "permutation_cycle", "status": "training", "step": 1395, "tokens_per_sec_per_gpu": 105698.042658, "vram_gb": 20.709735} {"adam_lr": 2.5302906670788462e-05, "dt_ms": 153.984785, "global_tokens_per_sec": 106400.122528, "grad_norm": 30.0, "loss": 2.369645, "lr_stage": "cosine_decay", "mfu": 61.120331, "stage_name": "permutation_cycle", "status": "training", "step": 1396, "tokens_per_sec_per_gpu": 106400.122528, "vram_gb": 20.709735} {"adam_lr": 2.522706657837962e-05, "dt_ms": 155.869961, "global_tokens_per_sec": 105113.261834, "grad_norm": 32.75, "loss": 2.530354, "lr_stage": "cosine_decay", "mfu": 60.381109, "stage_name": "permutation_cycle", "status": "training", "step": 1397, "tokens_per_sec_per_gpu": 105113.261834, "vram_gb": 20.709735} {"adam_lr": 2.5151301948622237e-05, "dt_ms": 154.530287, "global_tokens_per_sec": 106024.52335, "grad_norm": 30.375, "loss": 2.688571, "lr_stage": "cosine_decay", "mfu": 60.904573, "stage_name": "permutation_cycle", "status": "training", "step": 1398, "tokens_per_sec_per_gpu": 106024.52335, "vram_gb": 20.709735} {"adam_lr": 2.507561301230849e-05, "dt_ms": 154.597044, "global_tokens_per_sec": 105978.740453, "grad_norm": 35.25, "loss": 2.423024, "lr_stage": "cosine_decay", "mfu": 60.878273, "stage_name": "permutation_cycle", "status": "training", "step": 1399, "tokens_per_sec_per_gpu": 105978.740453, "vram_gb": 20.709735} {"adam_lr": 2.500000000000001e-05, "dt_ms": 155.317307, "global_tokens_per_sec": 105487.278702, "grad_norm": 44.0, "loss": 2.94775, "lr_stage": "cosine_decay", "mfu": 60.595959, "stage_name": "permutation_cycle", "status": "training", "step": 1400, "tokens_per_sec_per_gpu": 105487.278702, "vram_gb": 20.709735} {"adam_lr": 2.4924463142027127e-05, "dt_ms": 155.059099, "global_tokens_per_sec": 105662.938098, "grad_norm": 30.375, "loss": 2.368729, "lr_stage": "cosine_decay", "mfu": 60.696864, "stage_name": "permutation_cycle", "status": "training", "step": 1401, "tokens_per_sec_per_gpu": 105662.938098, "vram_gb": 20.709735} {"adam_lr": 2.4849002668488245e-05, "dt_ms": 155.173063, "global_tokens_per_sec": 105585.335843, "grad_norm": 30.75, "loss": 1.872898, "lr_stage": "cosine_decay", "mfu": 60.652286, "stage_name": "permutation_cycle", "status": "training", "step": 1402, "tokens_per_sec_per_gpu": 105585.335843, "vram_gb": 20.709735} {"adam_lr": 2.4773618809249042e-05, "dt_ms": 155.61223, "global_tokens_per_sec": 105287.354139, "grad_norm": 31.125, "loss": 2.778573, "lr_stage": "cosine_decay", "mfu": 60.481114, "stage_name": "permutation_cycle", "status": "training", "step": 1403, "tokens_per_sec_per_gpu": 105287.354139, "vram_gb": 20.709735} {"adam_lr": 2.469831179394182e-05, "dt_ms": 155.259132, "global_tokens_per_sec": 105526.803791, "grad_norm": 27.375, "loss": 1.858223, "lr_stage": "cosine_decay", "mfu": 60.618663, "stage_name": "permutation_cycle", "status": "training", "step": 1404, "tokens_per_sec_per_gpu": 105526.803791, "vram_gb": 20.709735} {"adam_lr": 2.4623081851964806e-05, "dt_ms": 155.967951, "global_tokens_per_sec": 105047.22229, "grad_norm": 24.25, "loss": 1.780029, "lr_stage": "cosine_decay", "mfu": 60.343173, "stage_name": "permutation_cycle", "status": "training", "step": 1405, "tokens_per_sec_per_gpu": 105047.22229, "vram_gb": 20.709735} {"adam_lr": 2.4547929212481435e-05, "dt_ms": 157.066822, "global_tokens_per_sec": 104312.290692, "grad_norm": 19.25, "loss": 1.645191, "lr_stage": "cosine_decay", "mfu": 59.921, "stage_name": "permutation_cycle", "status": "training", "step": 1406, "tokens_per_sec_per_gpu": 104312.290692, "vram_gb": 20.709735} {"adam_lr": 2.447285410441972e-05, "dt_ms": 155.953407, "global_tokens_per_sec": 105057.018535, "grad_norm": 27.625, "loss": 2.062057, "lr_stage": "cosine_decay", "mfu": 60.348801, "stage_name": "permutation_cycle", "status": "training", "step": 1407, "tokens_per_sec_per_gpu": 105057.018535, "vram_gb": 20.709735} {"adam_lr": 2.4397856756471432e-05, "dt_ms": 155.902624, "global_tokens_per_sec": 105091.239428, "grad_norm": 24.5, "loss": 2.738448, "lr_stage": "cosine_decay", "mfu": 60.368458, "stage_name": "permutation_cycle", "status": "training", "step": 1408, "tokens_per_sec_per_gpu": 105091.239428, "vram_gb": 20.709735} {"adam_lr": 2.432293739709151e-05, "dt_ms": 155.512571, "global_tokens_per_sec": 105354.826683, "grad_norm": 35.5, "loss": 2.756444, "lr_stage": "cosine_decay", "mfu": 60.519873, "stage_name": "permutation_cycle", "status": "training", "step": 1409, "tokens_per_sec_per_gpu": 105354.826683, "vram_gb": 20.709735} {"adam_lr": 2.4248096254497288e-05, "dt_ms": 155.514002, "global_tokens_per_sec": 105353.857566, "grad_norm": 29.875, "loss": 1.674381, "lr_stage": "cosine_decay", "mfu": 60.519316, "stage_name": "permutation_cycle", "status": "training", "step": 1410, "tokens_per_sec_per_gpu": 105353.857566, "vram_gb": 20.709735} {"adam_lr": 2.417333355666791e-05, "dt_ms": 155.581951, "global_tokens_per_sec": 105307.845028, "grad_norm": 29.375, "loss": 2.969515, "lr_stage": "cosine_decay", "mfu": 60.492885, "stage_name": "permutation_cycle", "status": "training", "step": 1411, "tokens_per_sec_per_gpu": 105307.845028, "vram_gb": 20.709735} {"adam_lr": 2.4098649531343497e-05, "dt_ms": 154.821873, "global_tokens_per_sec": 105824.840593, "grad_norm": 29.0, "loss": 1.74415, "lr_stage": "cosine_decay", "mfu": 60.789867, "stage_name": "permutation_cycle", "status": "training", "step": 1412, "tokens_per_sec_per_gpu": 105824.840593, "vram_gb": 20.709735} {"adam_lr": 2.4024044406024548e-05, "dt_ms": 156.085491, "global_tokens_per_sec": 104968.116358, "grad_norm": 27.5, "loss": 2.317312, "lr_stage": "cosine_decay", "mfu": 60.297732, "stage_name": "permutation_cycle", "status": "training", "step": 1413, "tokens_per_sec_per_gpu": 104968.116358, "vram_gb": 20.709735} {"adam_lr": 2.39495184079712e-05, "dt_ms": 154.702663, "global_tokens_per_sec": 105906.38608, "grad_norm": 32.75, "loss": 3.27824, "lr_stage": "cosine_decay", "mfu": 60.83671, "stage_name": "permutation_cycle", "status": "training", "step": 1414, "tokens_per_sec_per_gpu": 105906.38608, "vram_gb": 20.709735} {"adam_lr": 2.3875071764202563e-05, "dt_ms": 154.617071, "global_tokens_per_sec": 105965.013294, "grad_norm": 34.25, "loss": 2.530668, "lr_stage": "cosine_decay", "mfu": 60.870388, "stage_name": "permutation_cycle", "status": "training", "step": 1415, "tokens_per_sec_per_gpu": 105965.013294, "vram_gb": 20.709735} {"adam_lr": 2.3800704701496053e-05, "dt_ms": 154.965878, "global_tokens_per_sec": 105726.500962, "grad_norm": 23.75, "loss": 1.744818, "lr_stage": "cosine_decay", "mfu": 60.733377, "stage_name": "permutation_cycle", "status": "training", "step": 1416, "tokens_per_sec_per_gpu": 105726.500962, "vram_gb": 20.709735} {"adam_lr": 2.3726417446386624e-05, "dt_ms": 155.90477, "global_tokens_per_sec": 105089.793024, "grad_norm": 30.125, "loss": 2.906139, "lr_stage": "cosine_decay", "mfu": 60.367628, "stage_name": "permutation_cycle", "status": "training", "step": 1417, "tokens_per_sec_per_gpu": 105089.793024, "vram_gb": 20.709735} {"adam_lr": 2.3652210225166122e-05, "dt_ms": 155.842781, "global_tokens_per_sec": 105131.594084, "grad_norm": 37.5, "loss": 2.940584, "lr_stage": "cosine_decay", "mfu": 60.39164, "stage_name": "permutation_cycle", "status": "training", "step": 1418, "tokens_per_sec_per_gpu": 105131.594084, "vram_gb": 20.709735} {"adam_lr": 2.357808326388265e-05, "dt_ms": 154.793739, "global_tokens_per_sec": 105844.074005, "grad_norm": 22.375, "loss": 1.558811, "lr_stage": "cosine_decay", "mfu": 60.800916, "stage_name": "permutation_cycle", "status": "training", "step": 1419, "tokens_per_sec_per_gpu": 105844.074005, "vram_gb": 20.709735} {"adam_lr": 2.350403678833976e-05, "dt_ms": 156.802416, "global_tokens_per_sec": 104488.186049, "grad_norm": 25.375, "loss": 2.738092, "lr_stage": "cosine_decay", "mfu": 60.022041, "stage_name": "permutation_cycle", "status": "training", "step": 1420, "tokens_per_sec_per_gpu": 104488.186049, "vram_gb": 20.709735} {"adam_lr": 2.343007102409585e-05, "dt_ms": 155.89571, "global_tokens_per_sec": 105095.900336, "grad_norm": 28.375, "loss": 2.232842, "lr_stage": "cosine_decay", "mfu": 60.371136, "stage_name": "permutation_cycle", "status": "training", "step": 1421, "tokens_per_sec_per_gpu": 105095.900336, "vram_gb": 20.709735} {"adam_lr": 2.33561861964635e-05, "dt_ms": 154.819012, "global_tokens_per_sec": 105826.796214, "grad_norm": 26.75, "loss": 2.246099, "lr_stage": "cosine_decay", "mfu": 60.790991, "stage_name": "permutation_cycle", "status": "training", "step": 1422, "tokens_per_sec_per_gpu": 105826.796214, "vram_gb": 20.709735} {"adam_lr": 2.3282382530508685e-05, "dt_ms": 155.505896, "global_tokens_per_sec": 105359.349465, "grad_norm": 32.75, "loss": 2.691187, "lr_stage": "cosine_decay", "mfu": 60.522471, "stage_name": "permutation_cycle", "status": "training", "step": 1423, "tokens_per_sec_per_gpu": 105359.349465, "vram_gb": 20.709735} {"adam_lr": 2.3208660251050158e-05, "dt_ms": 155.256271, "global_tokens_per_sec": 105528.748412, "grad_norm": 32.5, "loss": 2.252412, "lr_stage": "cosine_decay", "mfu": 60.61978, "stage_name": "permutation_cycle", "status": "training", "step": 1424, "tokens_per_sec_per_gpu": 105528.748412, "vram_gb": 20.709735} {"adam_lr": 2.3135019582658802e-05, "dt_ms": 155.970573, "global_tokens_per_sec": 105045.455948, "grad_norm": 22.875, "loss": 2.284328, "lr_stage": "cosine_decay", "mfu": 60.342159, "stage_name": "permutation_cycle", "status": "training", "step": 1425, "tokens_per_sec_per_gpu": 105045.455948, "vram_gb": 20.709735} {"adam_lr": 2.3061460749656844e-05, "dt_ms": 154.620409, "global_tokens_per_sec": 105962.725779, "grad_norm": 41.25, "loss": 2.776658, "lr_stage": "cosine_decay", "mfu": 60.869074, "stage_name": "permutation_cycle", "status": "training", "step": 1426, "tokens_per_sec_per_gpu": 105962.725779, "vram_gb": 20.709735} {"adam_lr": 2.298798397611725e-05, "dt_ms": 154.762745, "global_tokens_per_sec": 105865.271453, "grad_norm": 26.375, "loss": 2.317079, "lr_stage": "cosine_decay", "mfu": 60.813092, "stage_name": "permutation_cycle", "status": "training", "step": 1427, "tokens_per_sec_per_gpu": 105865.271453, "vram_gb": 20.709735} {"adam_lr": 2.2914589485863014e-05, "dt_ms": 154.040575, "global_tokens_per_sec": 106361.586855, "grad_norm": 25.875, "loss": 2.007408, "lr_stage": "cosine_decay", "mfu": 61.098195, "stage_name": "permutation_cycle", "status": "training", "step": 1428, "tokens_per_sec_per_gpu": 106361.586855, "vram_gb": 20.709735} {"adam_lr": 2.284127750246646e-05, "dt_ms": 154.637098, "global_tokens_per_sec": 105951.289689, "grad_norm": 40.75, "loss": 2.346318, "lr_stage": "cosine_decay", "mfu": 60.862504, "stage_name": "permutation_cycle", "status": "training", "step": 1429, "tokens_per_sec_per_gpu": 105951.289689, "vram_gb": 20.709735} {"adam_lr": 2.2768048249248648e-05, "dt_ms": 155.985355, "global_tokens_per_sec": 105035.501316, "grad_norm": 24.875, "loss": 2.760166, "lr_stage": "cosine_decay", "mfu": 60.33644, "stage_name": "permutation_cycle", "status": "training", "step": 1430, "tokens_per_sec_per_gpu": 105035.501316, "vram_gb": 20.709735} {"adam_lr": 2.2694901949278553e-05, "dt_ms": 155.087948, "global_tokens_per_sec": 105643.283231, "grad_norm": 34.75, "loss": 2.232662, "lr_stage": "cosine_decay", "mfu": 60.685574, "stage_name": "permutation_cycle", "status": "training", "step": 1431, "tokens_per_sec_per_gpu": 105643.283231, "vram_gb": 20.709735} {"adam_lr": 2.2621838825372493e-05, "dt_ms": 155.476332, "global_tokens_per_sec": 105379.3836, "grad_norm": 25.375, "loss": 1.907978, "lr_stage": "cosine_decay", "mfu": 60.53398, "stage_name": "permutation_cycle", "status": "training", "step": 1432, "tokens_per_sec_per_gpu": 105379.3836, "vram_gb": 20.709735} {"adam_lr": 2.2548859100093407e-05, "dt_ms": 154.955864, "global_tokens_per_sec": 105733.333235, "grad_norm": 42.25, "loss": 3.15769, "lr_stage": "cosine_decay", "mfu": 60.737302, "stage_name": "permutation_cycle", "status": "training", "step": 1433, "tokens_per_sec_per_gpu": 105733.333235, "vram_gb": 20.709735} {"adam_lr": 2.247596299575022e-05, "dt_ms": 155.802488, "global_tokens_per_sec": 105158.782609, "grad_norm": 18.25, "loss": 2.417629, "lr_stage": "cosine_decay", "mfu": 60.407258, "stage_name": "permutation_cycle", "status": "training", "step": 1434, "tokens_per_sec_per_gpu": 105158.782609, "vram_gb": 20.709735} {"adam_lr": 2.2403150734397094e-05, "dt_ms": 154.735565, "global_tokens_per_sec": 105883.866972, "grad_norm": 35.25, "loss": 2.258291, "lr_stage": "cosine_decay", "mfu": 60.823774, "stage_name": "permutation_cycle", "status": "training", "step": 1435, "tokens_per_sec_per_gpu": 105883.866972, "vram_gb": 20.709735} {"adam_lr": 2.23304225378328e-05, "dt_ms": 157.121181, "global_tokens_per_sec": 104276.201622, "grad_norm": 31.625, "loss": 2.625149, "lr_stage": "cosine_decay", "mfu": 59.900269, "stage_name": "permutation_cycle", "status": "training", "step": 1436, "tokens_per_sec_per_gpu": 104276.201622, "vram_gb": 20.709735} {"adam_lr": 2.2257778627600045e-05, "dt_ms": 155.9515, "global_tokens_per_sec": 105058.303424, "grad_norm": 19.75, "loss": 2.000849, "lr_stage": "cosine_decay", "mfu": 60.349539, "stage_name": "permutation_cycle", "status": "training", "step": 1437, "tokens_per_sec_per_gpu": 105058.303424, "vram_gb": 20.709735} {"adam_lr": 2.218521922498476e-05, "dt_ms": 154.767513, "global_tokens_per_sec": 105862.009754, "grad_norm": 34.75, "loss": 2.017676, "lr_stage": "cosine_decay", "mfu": 60.811219, "stage_name": "permutation_cycle", "status": "training", "step": 1438, "tokens_per_sec_per_gpu": 105862.009754, "vram_gb": 20.709735} {"adam_lr": 2.2112744551015498e-05, "dt_ms": 153.398991, "global_tokens_per_sec": 106806.43942, "grad_norm": 37.5, "loss": 1.619797, "lr_stage": "cosine_decay", "mfu": 61.353735, "stage_name": "permutation_cycle", "status": "training", "step": 1439, "tokens_per_sec_per_gpu": 106806.43942, "vram_gb": 20.709735} {"adam_lr": 2.2040354826462668e-05, "dt_ms": 155.977011, "global_tokens_per_sec": 105041.120635, "grad_norm": 31.25, "loss": 2.525293, "lr_stage": "cosine_decay", "mfu": 60.339668, "stage_name": "permutation_cycle", "status": "training", "step": 1440, "tokens_per_sec_per_gpu": 105041.120635, "vram_gb": 20.709735} {"adam_lr": 2.1968050271837925e-05, "dt_ms": 154.540062, "global_tokens_per_sec": 106017.816954, "grad_norm": 28.5, "loss": 1.545722, "lr_stage": "cosine_decay", "mfu": 60.90072, "stage_name": "permutation_cycle", "status": "training", "step": 1441, "tokens_per_sec_per_gpu": 106017.816954, "vram_gb": 20.709735} {"adam_lr": 2.1895831107393484e-05, "dt_ms": 155.222416, "global_tokens_per_sec": 105551.765204, "grad_norm": 50.0, "loss": 2.975808, "lr_stage": "cosine_decay", "mfu": 60.633002, "stage_name": "permutation_cycle", "status": "training", "step": 1442, "tokens_per_sec_per_gpu": 105551.765204, "vram_gb": 20.709735} {"adam_lr": 2.182369755312143e-05, "dt_ms": 155.468941, "global_tokens_per_sec": 105384.393324, "grad_norm": 32.25, "loss": 2.386269, "lr_stage": "cosine_decay", "mfu": 60.536857, "stage_name": "permutation_cycle", "status": "training", "step": 1443, "tokens_per_sec_per_gpu": 105384.393324, "vram_gb": 20.709735} {"adam_lr": 2.1751649828753106e-05, "dt_ms": 157.165289, "global_tokens_per_sec": 104246.937171, "grad_norm": 17.75, "loss": 2.575214, "lr_stage": "cosine_decay", "mfu": 59.883459, "stage_name": "permutation_cycle", "status": "training", "step": 1444, "tokens_per_sec_per_gpu": 104246.937171, "vram_gb": 20.709735} {"adam_lr": 2.167968815375837e-05, "dt_ms": 154.320955, "global_tokens_per_sec": 106168.342275, "grad_norm": 28.5, "loss": 1.37803, "lr_stage": "cosine_decay", "mfu": 60.987188, "stage_name": "permutation_cycle", "status": "training", "step": 1445, "tokens_per_sec_per_gpu": 106168.342275, "vram_gb": 20.709735} {"adam_lr": 2.160781274734495e-05, "dt_ms": 156.356573, "global_tokens_per_sec": 104786.128748, "grad_norm": 26.875, "loss": 2.112023, "lr_stage": "cosine_decay", "mfu": 60.193191, "stage_name": "permutation_cycle", "status": "training", "step": 1446, "tokens_per_sec_per_gpu": 104786.128748, "vram_gb": 20.709735} {"adam_lr": 2.153602382845779e-05, "dt_ms": 154.520988, "global_tokens_per_sec": 106030.903393, "grad_norm": 31.125, "loss": 2.582843, "lr_stage": "cosine_decay", "mfu": 60.908238, "stage_name": "permutation_cycle", "status": "training", "step": 1447, "tokens_per_sec_per_gpu": 106030.903393, "vram_gb": 20.709735} {"adam_lr": 2.1464321615778422e-05, "dt_ms": 154.991388, "global_tokens_per_sec": 105709.098922, "grad_norm": 22.5, "loss": 2.182917, "lr_stage": "cosine_decay", "mfu": 60.723381, "stage_name": "permutation_cycle", "status": "training", "step": 1448, "tokens_per_sec_per_gpu": 105709.098922, "vram_gb": 20.709735} {"adam_lr": 2.13927063277242e-05, "dt_ms": 156.190872, "global_tokens_per_sec": 104897.295021, "grad_norm": 27.5, "loss": 1.917259, "lr_stage": "cosine_decay", "mfu": 60.257049, "stage_name": "permutation_cycle", "status": "training", "step": 1449, "tokens_per_sec_per_gpu": 104897.295021, "vram_gb": 20.709735} {"adam_lr": 2.132117818244771e-05, "dt_ms": 155.674219, "global_tokens_per_sec": 105245.429149, "grad_norm": 30.375, "loss": 2.422387, "lr_stage": "cosine_decay", "mfu": 60.457031, "stage_name": "permutation_cycle", "status": "training", "step": 1450, "tokens_per_sec_per_gpu": 105245.429149, "vram_gb": 20.709735} {"adam_lr": 2.124973739783609e-05, "dt_ms": 154.995918, "global_tokens_per_sec": 105706.009439, "grad_norm": 27.75, "loss": 2.17235, "lr_stage": "cosine_decay", "mfu": 60.721606, "stage_name": "permutation_cycle", "status": "training", "step": 1451, "tokens_per_sec_per_gpu": 105706.009439, "vram_gb": 20.709735} {"adam_lr": 2.117838419151034e-05, "dt_ms": 155.324697, "global_tokens_per_sec": 105482.259192, "grad_norm": 21.75, "loss": 2.351363, "lr_stage": "cosine_decay", "mfu": 60.593075, "stage_name": "permutation_cycle", "status": "training", "step": 1452, "tokens_per_sec_per_gpu": 105482.259192, "vram_gb": 20.709735} {"adam_lr": 2.110711878082474e-05, "dt_ms": 154.243946, "global_tokens_per_sec": 106221.348824, "grad_norm": 36.25, "loss": 1.977468, "lr_stage": "cosine_decay", "mfu": 61.017637, "stage_name": "permutation_cycle", "status": "training", "step": 1453, "tokens_per_sec_per_gpu": 106221.348824, "vram_gb": 20.709735} {"adam_lr": 2.103594138286607e-05, "dt_ms": 155.416727, "global_tokens_per_sec": 105419.798173, "grad_norm": 26.125, "loss": 3.097106, "lr_stage": "cosine_decay", "mfu": 60.557195, "stage_name": "permutation_cycle", "status": "training", "step": 1454, "tokens_per_sec_per_gpu": 105419.798173, "vram_gb": 20.709735} {"adam_lr": 2.0964852214453013e-05, "dt_ms": 156.048298, "global_tokens_per_sec": 104993.134961, "grad_norm": 28.0, "loss": 1.747217, "lr_stage": "cosine_decay", "mfu": 60.312103, "stage_name": "permutation_cycle", "status": "training", "step": 1455, "tokens_per_sec_per_gpu": 104993.134961, "vram_gb": 20.709735} {"adam_lr": 2.0893851492135537e-05, "dt_ms": 155.619621, "global_tokens_per_sec": 105282.353636, "grad_norm": 32.75, "loss": 1.74367, "lr_stage": "cosine_decay", "mfu": 60.478242, "stage_name": "permutation_cycle", "status": "training", "step": 1456, "tokens_per_sec_per_gpu": 105282.353636, "vram_gb": 20.709735} {"adam_lr": 2.0822939432194133e-05, "dt_ms": 156.409979, "global_tokens_per_sec": 104750.34981, "grad_norm": 29.875, "loss": 2.552834, "lr_stage": "cosine_decay", "mfu": 60.172638, "stage_name": "permutation_cycle", "status": "training", "step": 1457, "tokens_per_sec_per_gpu": 104750.34981, "vram_gb": 20.709735} {"adam_lr": 2.0752116250639225e-05, "dt_ms": 155.508995, "global_tokens_per_sec": 105357.249554, "grad_norm": 19.0, "loss": 2.211391, "lr_stage": "cosine_decay", "mfu": 60.521265, "stage_name": "permutation_cycle", "status": "training", "step": 1458, "tokens_per_sec_per_gpu": 105357.249554, "vram_gb": 20.709735} {"adam_lr": 2.0681382163210532e-05, "dt_ms": 155.701876, "global_tokens_per_sec": 105226.734924, "grad_norm": 19.5, "loss": 1.970104, "lr_stage": "cosine_decay", "mfu": 60.446292, "stage_name": "permutation_cycle", "status": "training", "step": 1459, "tokens_per_sec_per_gpu": 105226.734924, "vram_gb": 20.709735} {"adam_lr": 2.061073738537635e-05, "dt_ms": 157.163382, "global_tokens_per_sec": 104248.202321, "grad_norm": 29.25, "loss": 3.125309, "lr_stage": "cosine_decay", "mfu": 59.884185, "stage_name": "permutation_cycle", "status": "training", "step": 1460, "tokens_per_sec_per_gpu": 104248.202321, "vram_gb": 20.709735} {"adam_lr": 2.0540182132332887e-05, "dt_ms": 154.83737, "global_tokens_per_sec": 105814.248903, "grad_norm": 13.625, "loss": 2.020901, "lr_stage": "cosine_decay", "mfu": 60.783783, "stage_name": "permutation_cycle", "status": "training", "step": 1461, "tokens_per_sec_per_gpu": 105814.248903, "vram_gb": 20.709735} {"adam_lr": 2.0469716619003725e-05, "dt_ms": 156.591177, "global_tokens_per_sec": 104629.138852, "grad_norm": 30.125, "loss": 2.816592, "lr_stage": "cosine_decay", "mfu": 60.10301, "stage_name": "permutation_cycle", "status": "training", "step": 1462, "tokens_per_sec_per_gpu": 104629.138852, "vram_gb": 20.709735} {"adam_lr": 2.0399341060039022e-05, "dt_ms": 156.179905, "global_tokens_per_sec": 104904.661112, "grad_norm": 26.0, "loss": 2.427374, "lr_stage": "cosine_decay", "mfu": 60.261281, "stage_name": "permutation_cycle", "status": "training", "step": 1463, "tokens_per_sec_per_gpu": 104904.661112, "vram_gb": 20.709735} {"adam_lr": 2.0329055669814934e-05, "dt_ms": 156.633139, "global_tokens_per_sec": 104601.108939, "grad_norm": 40.5, "loss": 2.322843, "lr_stage": "cosine_decay", "mfu": 60.086909, "stage_name": "permutation_cycle", "status": "training", "step": 1464, "tokens_per_sec_per_gpu": 104601.108939, "vram_gb": 20.709735} {"adam_lr": 2.0258860662432942e-05, "dt_ms": 155.897856, "global_tokens_per_sec": 105094.453803, "grad_norm": 30.625, "loss": 2.562593, "lr_stage": "cosine_decay", "mfu": 60.370305, "stage_name": "permutation_cycle", "status": "training", "step": 1465, "tokens_per_sec_per_gpu": 105094.453803, "vram_gb": 20.709735} {"adam_lr": 2.0188756251719203e-05, "dt_ms": 155.258894, "global_tokens_per_sec": 105526.96584, "grad_norm": 21.875, "loss": 1.583599, "lr_stage": "cosine_decay", "mfu": 60.618756, "stage_name": "permutation_cycle", "status": "training", "step": 1466, "tokens_per_sec_per_gpu": 105526.96584, "vram_gb": 20.709735} {"adam_lr": 2.0118742651223944e-05, "dt_ms": 155.576944, "global_tokens_per_sec": 105311.234054, "grad_norm": 34.75, "loss": 2.216681, "lr_stage": "cosine_decay", "mfu": 60.494832, "stage_name": "permutation_cycle", "status": "training", "step": 1467, "tokens_per_sec_per_gpu": 105311.234054, "vram_gb": 20.709735} {"adam_lr": 2.0048820074220715e-05, "dt_ms": 156.644821, "global_tokens_per_sec": 104593.307828, "grad_norm": 26.25, "loss": 2.881401, "lr_stage": "cosine_decay", "mfu": 60.082427, "stage_name": "permutation_cycle", "status": "training", "step": 1468, "tokens_per_sec_per_gpu": 104593.307828, "vram_gb": 20.709735} {"adam_lr": 1.9978988733705807e-05, "dt_ms": 157.235384, "global_tokens_per_sec": 104200.464199, "grad_norm": 21.0, "loss": 2.043981, "lr_stage": "cosine_decay", "mfu": 59.856763, "stage_name": "permutation_cycle", "status": "training", "step": 1469, "tokens_per_sec_per_gpu": 104200.464199, "vram_gb": 20.709735} {"adam_lr": 1.9909248842397584e-05, "dt_ms": 155.945301, "global_tokens_per_sec": 105062.47953, "grad_norm": 21.125, "loss": 1.609934, "lr_stage": "cosine_decay", "mfu": 60.351938, "stage_name": "permutation_cycle", "status": "training", "step": 1470, "tokens_per_sec_per_gpu": 105062.47953, "vram_gb": 20.709735} {"adam_lr": 1.9839600612735874e-05, "dt_ms": 155.551195, "global_tokens_per_sec": 105328.666776, "grad_norm": 31.375, "loss": 3.142442, "lr_stage": "cosine_decay", "mfu": 60.504846, "stage_name": "permutation_cycle", "status": "training", "step": 1471, "tokens_per_sec_per_gpu": 105328.666776, "vram_gb": 20.709735} {"adam_lr": 1.977004425688126e-05, "dt_ms": 154.54793, "global_tokens_per_sec": 106012.41974, "grad_norm": 37.0, "loss": 3.04472, "lr_stage": "cosine_decay", "mfu": 60.89762, "stage_name": "permutation_cycle", "status": "training", "step": 1472, "tokens_per_sec_per_gpu": 106012.41974, "vram_gb": 20.709735} {"adam_lr": 1.9700579986714455e-05, "dt_ms": 156.121492, "global_tokens_per_sec": 104943.910986, "grad_norm": 26.75, "loss": 2.443366, "lr_stage": "cosine_decay", "mfu": 60.283827, "stage_name": "permutation_cycle", "status": "training", "step": 1473, "tokens_per_sec_per_gpu": 104943.910986, "vram_gb": 20.709735} {"adam_lr": 1.9631208013835678e-05, "dt_ms": 155.549765, "global_tokens_per_sec": 105329.63543, "grad_norm": 40.25, "loss": 2.523513, "lr_stage": "cosine_decay", "mfu": 60.505402, "stage_name": "permutation_cycle", "status": "training", "step": 1474, "tokens_per_sec_per_gpu": 105329.63543, "vram_gb": 20.709735} {"adam_lr": 1.9561928549563968e-05, "dt_ms": 156.504393, "global_tokens_per_sec": 104687.1575, "grad_norm": 34.0, "loss": 2.487806, "lr_stage": "cosine_decay", "mfu": 60.136338, "stage_name": "permutation_cycle", "status": "training", "step": 1475, "tokens_per_sec_per_gpu": 104687.1575, "vram_gb": 20.709735} {"adam_lr": 1.9492741804936622e-05, "dt_ms": 156.071663, "global_tokens_per_sec": 104977.416754, "grad_norm": 27.5, "loss": 1.80747, "lr_stage": "cosine_decay", "mfu": 60.303074, "stage_name": "permutation_cycle", "status": "training", "step": 1476, "tokens_per_sec_per_gpu": 104977.416754, "vram_gb": 20.709735} {"adam_lr": 1.9423647990708448e-05, "dt_ms": 156.450033, "global_tokens_per_sec": 104723.531636, "grad_norm": 26.5, "loss": 3.102347, "lr_stage": "cosine_decay", "mfu": 60.157233, "stage_name": "permutation_cycle", "status": "training", "step": 1477, "tokens_per_sec_per_gpu": 104723.531636, "vram_gb": 20.709735} {"adam_lr": 1.9354647317351188e-05, "dt_ms": 156.215668, "global_tokens_per_sec": 104880.645064, "grad_norm": 28.125, "loss": 2.666189, "lr_stage": "cosine_decay", "mfu": 60.247485, "stage_name": "permutation_cycle", "status": "training", "step": 1478, "tokens_per_sec_per_gpu": 104880.645064, "vram_gb": 20.709735} {"adam_lr": 1.9285739995052836e-05, "dt_ms": 156.696796, "global_tokens_per_sec": 104558.614947, "grad_norm": 35.75, "loss": 2.043716, "lr_stage": "cosine_decay", "mfu": 60.062498, "stage_name": "permutation_cycle", "status": "training", "step": 1479, "tokens_per_sec_per_gpu": 104558.614947, "vram_gb": 20.709735} {"adam_lr": 1.9216926233717085e-05, "dt_ms": 154.42872, "global_tokens_per_sec": 106094.254681, "grad_norm": 15.625, "loss": 1.405124, "lr_stage": "cosine_decay", "mfu": 60.944629, "stage_name": "permutation_cycle", "status": "training", "step": 1480, "tokens_per_sec_per_gpu": 106094.254681, "vram_gb": 20.709735} {"adam_lr": 1.9148206242962575e-05, "dt_ms": 155.98464, "global_tokens_per_sec": 105035.982948, "grad_norm": 26.0, "loss": 2.13186, "lr_stage": "cosine_decay", "mfu": 60.336717, "stage_name": "permutation_cycle", "status": "training", "step": 1481, "tokens_per_sec_per_gpu": 105035.982948, "vram_gb": 20.709735} {"adam_lr": 1.9079580232122303e-05, "dt_ms": 155.495405, "global_tokens_per_sec": 105366.45748, "grad_norm": 15.5, "loss": 1.983775, "lr_stage": "cosine_decay", "mfu": 60.526554, "stage_name": "permutation_cycle", "status": "training", "step": 1482, "tokens_per_sec_per_gpu": 105366.45748, "vram_gb": 20.709735} {"adam_lr": 1.901104841024301e-05, "dt_ms": 155.413866, "global_tokens_per_sec": 105421.738852, "grad_norm": 18.125, "loss": 1.720009, "lr_stage": "cosine_decay", "mfu": 60.55831, "stage_name": "permutation_cycle", "status": "training", "step": 1483, "tokens_per_sec_per_gpu": 105421.738852, "vram_gb": 20.709735} {"adam_lr": 1.8942610986084486e-05, "dt_ms": 155.321598, "global_tokens_per_sec": 105484.36409, "grad_norm": 33.0, "loss": 1.934883, "lr_stage": "cosine_decay", "mfu": 60.594284, "stage_name": "permutation_cycle", "status": "training", "step": 1484, "tokens_per_sec_per_gpu": 105484.36409, "vram_gb": 20.709735} {"adam_lr": 1.887426816811903e-05, "dt_ms": 157.395601, "global_tokens_per_sec": 104094.395698, "grad_norm": 26.875, "loss": 2.714689, "lr_stage": "cosine_decay", "mfu": 59.795833, "stage_name": "permutation_cycle", "status": "training", "step": 1485, "tokens_per_sec_per_gpu": 104094.395698, "vram_gb": 20.709735} {"adam_lr": 1.8806020164530702e-05, "dt_ms": 154.131889, "global_tokens_per_sec": 106298.573707, "grad_norm": 31.75, "loss": 2.564569, "lr_stage": "cosine_decay", "mfu": 61.061998, "stage_name": "permutation_cycle", "status": "training", "step": 1486, "tokens_per_sec_per_gpu": 106298.573707, "vram_gb": 20.709735} {"adam_lr": 1.8737867183214757e-05, "dt_ms": 153.989077, "global_tokens_per_sec": 106397.157254, "grad_norm": 20.75, "loss": 1.916843, "lr_stage": "cosine_decay", "mfu": 61.118628, "stage_name": "permutation_cycle", "status": "training", "step": 1487, "tokens_per_sec_per_gpu": 106397.157254, "vram_gb": 20.709735} {"adam_lr": 1.866980943177699e-05, "dt_ms": 155.113935, "global_tokens_per_sec": 105625.583867, "grad_norm": 23.125, "loss": 2.615222, "lr_stage": "cosine_decay", "mfu": 60.675406, "stage_name": "permutation_cycle", "status": "training", "step": 1488, "tokens_per_sec_per_gpu": 105625.583867, "vram_gb": 20.709735} {"adam_lr": 1.860184711753311e-05, "dt_ms": 157.098055, "global_tokens_per_sec": 104291.552253, "grad_norm": 25.875, "loss": 2.593292, "lr_stage": "cosine_decay", "mfu": 59.909087, "stage_name": "permutation_cycle", "status": "training", "step": 1489, "tokens_per_sec_per_gpu": 104291.552253, "vram_gb": 20.709735} {"adam_lr": 1.8533980447508137e-05, "dt_ms": 155.348539, "global_tokens_per_sec": 105466.070478, "grad_norm": 37.0, "loss": 3.503531, "lr_stage": "cosine_decay", "mfu": 60.583776, "stage_name": "permutation_cycle", "status": "training", "step": 1490, "tokens_per_sec_per_gpu": 105466.070478, "vram_gb": 20.709735} {"adam_lr": 1.8466209628435703e-05, "dt_ms": 155.574083, "global_tokens_per_sec": 105313.170738, "grad_norm": 29.0, "loss": 2.03318, "lr_stage": "cosine_decay", "mfu": 60.495944, "stage_name": "permutation_cycle", "status": "training", "step": 1491, "tokens_per_sec_per_gpu": 105313.170738, "vram_gb": 20.709735} {"adam_lr": 1.8398534866757454e-05, "dt_ms": 155.451059, "global_tokens_per_sec": 105396.515594, "grad_norm": 31.875, "loss": 2.461546, "lr_stage": "cosine_decay", "mfu": 60.543821, "stage_name": "permutation_cycle", "status": "training", "step": 1492, "tokens_per_sec_per_gpu": 105396.515594, "vram_gb": 20.709735} {"adam_lr": 1.83309563686225e-05, "dt_ms": 155.831337, "global_tokens_per_sec": 105139.314839, "grad_norm": 29.0, "loss": 2.576498, "lr_stage": "cosine_decay", "mfu": 60.396075, "stage_name": "permutation_cycle", "status": "training", "step": 1493, "tokens_per_sec_per_gpu": 105139.314839, "vram_gb": 20.709735} {"adam_lr": 1.8263474339886628e-05, "dt_ms": 156.723976, "global_tokens_per_sec": 104540.481961, "grad_norm": 30.75, "loss": 2.288469, "lr_stage": "cosine_decay", "mfu": 60.052082, "stage_name": "permutation_cycle", "status": "training", "step": 1494, "tokens_per_sec_per_gpu": 104540.481961, "vram_gb": 20.709735} {"adam_lr": 1.8196088986111797e-05, "dt_ms": 156.415224, "global_tokens_per_sec": 104746.837124, "grad_norm": 26.875, "loss": 2.752399, "lr_stage": "cosine_decay", "mfu": 60.17062, "stage_name": "permutation_cycle", "status": "training", "step": 1495, "tokens_per_sec_per_gpu": 104746.837124, "vram_gb": 20.709735} {"adam_lr": 1.8128800512565513e-05, "dt_ms": 156.263828, "global_tokens_per_sec": 104848.320821, "grad_norm": 41.75, "loss": 3.42055, "lr_stage": "cosine_decay", "mfu": 60.228917, "stage_name": "permutation_cycle", "status": "training", "step": 1496, "tokens_per_sec_per_gpu": 104848.320821, "vram_gb": 20.709735} {"adam_lr": 1.806160912422012e-05, "dt_ms": 155.979395, "global_tokens_per_sec": 105039.515054, "grad_norm": 27.625, "loss": 2.575099, "lr_stage": "cosine_decay", "mfu": 60.338746, "stage_name": "permutation_cycle", "status": "training", "step": 1497, "tokens_per_sec_per_gpu": 105039.515054, "vram_gb": 20.709735} {"adam_lr": 1.7994515025752217e-05, "dt_ms": 155.856609, "global_tokens_per_sec": 105122.266351, "grad_norm": 20.625, "loss": 1.583984, "lr_stage": "cosine_decay", "mfu": 60.386281, "stage_name": "permutation_cycle", "status": "training", "step": 1498, "tokens_per_sec_per_gpu": 105122.266351, "vram_gb": 20.709735} {"adam_lr": 1.7927518421542104e-05, "dt_ms": 157.34601, "global_tokens_per_sec": 104127.203342, "grad_norm": 29.5, "loss": 2.263539, "lr_stage": "cosine_decay", "mfu": 59.814679, "stage_name": "permutation_cycle", "status": "training", "step": 1499, "tokens_per_sec_per_gpu": 104127.203342, "vram_gb": 20.709735} {"adam_lr": 1.7860619515673033e-05, "dt_ms": 154.560089, "global_tokens_per_sec": 106004.07967, "grad_norm": 33.75, "loss": 1.789035, "lr_stage": "cosine_decay", "mfu": 60.892829, "stage_name": "permutation_cycle", "status": "training", "step": 1500, "tokens_per_sec_per_gpu": 106004.07967, "vram_gb": 20.709735} {"adam_lr": 1.7793818511930675e-05, "dt_ms": 155.683994, "global_tokens_per_sec": 105238.820949, "grad_norm": 36.25, "loss": 2.344912, "lr_stage": "cosine_decay", "mfu": 60.453235, "stage_name": "permutation_cycle", "status": "training", "step": 1501, "tokens_per_sec_per_gpu": 105238.820949, "vram_gb": 20.709735} {"adam_lr": 1.7727115613802465e-05, "dt_ms": 155.843019, "global_tokens_per_sec": 105131.433247, "grad_norm": 31.375, "loss": 3.117586, "lr_stage": "cosine_decay", "mfu": 60.391547, "stage_name": "permutation_cycle", "status": "training", "step": 1502, "tokens_per_sec_per_gpu": 105131.433247, "vram_gb": 20.709735} {"adam_lr": 1.7660511024477018e-05, "dt_ms": 154.572964, "global_tokens_per_sec": 105995.250439, "grad_norm": 27.25, "loss": 2.091022, "lr_stage": "cosine_decay", "mfu": 60.887757, "stage_name": "permutation_cycle", "status": "training", "step": 1503, "tokens_per_sec_per_gpu": 105995.250439, "vram_gb": 20.709735} {"adam_lr": 1.7594004946843456e-05, "dt_ms": 153.823376, "global_tokens_per_sec": 106511.769913, "grad_norm": 35.5, "loss": 1.901966, "lr_stage": "cosine_decay", "mfu": 61.184466, "stage_name": "permutation_cycle", "status": "training", "step": 1504, "tokens_per_sec_per_gpu": 106511.769913, "vram_gb": 20.709735} {"adam_lr": 1.7527597583490822e-05, "dt_ms": 155.080795, "global_tokens_per_sec": 105648.155657, "grad_norm": 34.25, "loss": 2.570385, "lr_stage": "cosine_decay", "mfu": 60.688373, "stage_name": "permutation_cycle", "status": "training", "step": 1505, "tokens_per_sec_per_gpu": 105648.155657, "vram_gb": 20.709735} {"adam_lr": 1.746128913670746e-05, "dt_ms": 155.016422, "global_tokens_per_sec": 105692.027721, "grad_norm": 30.5, "loss": 2.630796, "lr_stage": "cosine_decay", "mfu": 60.713574, "stage_name": "permutation_cycle", "status": "training", "step": 1506, "tokens_per_sec_per_gpu": 105692.027721, "vram_gb": 20.709735} {"adam_lr": 1.7395079808480386e-05, "dt_ms": 155.508757, "global_tokens_per_sec": 105357.411083, "grad_norm": 28.625, "loss": 2.308311, "lr_stage": "cosine_decay", "mfu": 60.521358, "stage_name": "permutation_cycle", "status": "training", "step": 1507, "tokens_per_sec_per_gpu": 105357.411083, "vram_gb": 20.709735} {"adam_lr": 1.7328969800494726e-05, "dt_ms": 156.000614, "global_tokens_per_sec": 105025.227545, "grad_norm": 43.0, "loss": 2.578733, "lr_stage": "cosine_decay", "mfu": 60.330539, "stage_name": "permutation_cycle", "status": "training", "step": 1508, "tokens_per_sec_per_gpu": 105025.227545, "vram_gb": 20.709735} {"adam_lr": 1.7262959314133016e-05, "dt_ms": 155.925512, "global_tokens_per_sec": 105075.813168, "grad_norm": 30.75, "loss": 2.242921, "lr_stage": "cosine_decay", "mfu": 60.359597, "stage_name": "permutation_cycle", "status": "training", "step": 1509, "tokens_per_sec_per_gpu": 105075.813168, "vram_gb": 20.709735} {"adam_lr": 1.7197048550474643e-05, "dt_ms": 156.347036, "global_tokens_per_sec": 104792.520417, "grad_norm": 24.625, "loss": 2.777452, "lr_stage": "cosine_decay", "mfu": 60.196863, "stage_name": "permutation_cycle", "status": "training", "step": 1510, "tokens_per_sec_per_gpu": 104792.520417, "vram_gb": 20.709735} {"adam_lr": 1.7131237710295206e-05, "dt_ms": 156.011581, "global_tokens_per_sec": 105017.844514, "grad_norm": 30.25, "loss": 2.675423, "lr_stage": "cosine_decay", "mfu": 60.326298, "stage_name": "permutation_cycle", "status": "training", "step": 1511, "tokens_per_sec_per_gpu": 105017.844514, "vram_gb": 20.709735} {"adam_lr": 1.7065526994065973e-05, "dt_ms": 154.408693, "global_tokens_per_sec": 106108.015348, "grad_norm": 32.25, "loss": 1.61393, "lr_stage": "cosine_decay", "mfu": 60.952534, "stage_name": "permutation_cycle", "status": "training", "step": 1512, "tokens_per_sec_per_gpu": 106108.015348, "vram_gb": 20.709735} {"adam_lr": 1.6999916601953163e-05, "dt_ms": 154.928684, "global_tokens_per_sec": 105751.882429, "grad_norm": 34.5, "loss": 2.073069, "lr_stage": "cosine_decay", "mfu": 60.747957, "stage_name": "permutation_cycle", "status": "training", "step": 1513, "tokens_per_sec_per_gpu": 105751.882429, "vram_gb": 20.709735} {"adam_lr": 1.6934406733817414e-05, "dt_ms": 154.05488, "global_tokens_per_sec": 106351.710409, "grad_norm": 21.25, "loss": 2.088407, "lr_stage": "cosine_decay", "mfu": 61.092521, "stage_name": "permutation_cycle", "status": "training", "step": 1514, "tokens_per_sec_per_gpu": 106351.710409, "vram_gb": 20.709735} {"adam_lr": 1.6868997589213136e-05, "dt_ms": 155.67565, "global_tokens_per_sec": 105244.462044, "grad_norm": 33.25, "loss": 1.669677, "lr_stage": "cosine_decay", "mfu": 60.456475, "stage_name": "permutation_cycle", "status": "training", "step": 1515, "tokens_per_sec_per_gpu": 105244.462044, "vram_gb": 20.709735} {"adam_lr": 1.680368936738792e-05, "dt_ms": 153.810263, "global_tokens_per_sec": 106520.850524, "grad_norm": 30.125, "loss": 1.963915, "lr_stage": "cosine_decay", "mfu": 61.189682, "stage_name": "permutation_cycle", "status": "training", "step": 1516, "tokens_per_sec_per_gpu": 106520.850524, "vram_gb": 20.709735} {"adam_lr": 1.6738482267281965e-05, "dt_ms": 156.176805, "global_tokens_per_sec": 104906.743021, "grad_norm": 29.625, "loss": 2.317726, "lr_stage": "cosine_decay", "mfu": 60.262477, "stage_name": "permutation_cycle", "status": "training", "step": 1517, "tokens_per_sec_per_gpu": 104906.743021, "vram_gb": 20.709735} {"adam_lr": 1.667337648752738e-05, "dt_ms": 158.092499, "global_tokens_per_sec": 103635.530632, "grad_norm": 29.75, "loss": 1.961189, "lr_stage": "cosine_decay", "mfu": 59.532243, "stage_name": "permutation_cycle", "status": "training", "step": 1518, "tokens_per_sec_per_gpu": 103635.530632, "vram_gb": 20.709735} {"adam_lr": 1.6608372226447677e-05, "dt_ms": 159.304142, "global_tokens_per_sec": 102847.294455, "grad_norm": 22.125, "loss": 2.423557, "lr_stage": "cosine_decay", "mfu": 59.07945, "stage_name": "permutation_cycle", "status": "training", "step": 1519, "tokens_per_sec_per_gpu": 102847.294455, "vram_gb": 20.709735} {"adam_lr": 1.6543469682057106e-05, "dt_ms": 159.996986, "global_tokens_per_sec": 102401.928747, "grad_norm": 41.0, "loss": 2.824163, "lr_stage": "cosine_decay", "mfu": 58.823615, "stage_name": "permutation_cycle", "status": "training", "step": 1520, "tokens_per_sec_per_gpu": 102401.928747, "vram_gb": 20.709735} {"adam_lr": 1.6478669052060048e-05, "dt_ms": 155.803204, "global_tokens_per_sec": 105158.29985, "grad_norm": 17.875, "loss": 1.937524, "lr_stage": "cosine_decay", "mfu": 60.40698, "stage_name": "permutation_cycle", "status": "training", "step": 1521, "tokens_per_sec_per_gpu": 105158.29985, "vram_gb": 20.709735} {"adam_lr": 1.6413970533850498e-05, "dt_ms": 156.057358, "global_tokens_per_sec": 104987.039587, "grad_norm": 31.75, "loss": 2.417737, "lr_stage": "cosine_decay", "mfu": 60.308602, "stage_name": "permutation_cycle", "status": "training", "step": 1522, "tokens_per_sec_per_gpu": 104987.039587, "vram_gb": 20.709735} {"adam_lr": 1.6349374324511345e-05, "dt_ms": 153.858185, "global_tokens_per_sec": 106487.672526, "grad_norm": 42.75, "loss": 1.766445, "lr_stage": "cosine_decay", "mfu": 61.170623, "stage_name": "permutation_cycle", "status": "training", "step": 1523, "tokens_per_sec_per_gpu": 106487.672526, "vram_gb": 20.709735} {"adam_lr": 1.6284880620813848e-05, "dt_ms": 154.784203, "global_tokens_per_sec": 105850.595393, "grad_norm": 27.5, "loss": 2.967269, "lr_stage": "cosine_decay", "mfu": 60.804662, "stage_name": "permutation_cycle", "status": "training", "step": 1524, "tokens_per_sec_per_gpu": 105850.595393, "vram_gb": 20.709735} {"adam_lr": 1.622048961921699e-05, "dt_ms": 154.365301, "global_tokens_per_sec": 106137.842377, "grad_norm": 21.125, "loss": 1.80233, "lr_stage": "cosine_decay", "mfu": 60.969667, "stage_name": "permutation_cycle", "status": "training", "step": 1525, "tokens_per_sec_per_gpu": 106137.842377, "vram_gb": 20.709735} {"adam_lr": 1.615620151586697e-05, "dt_ms": 154.443264, "global_tokens_per_sec": 106084.264052, "grad_norm": 32.5, "loss": 1.944653, "lr_stage": "cosine_decay", "mfu": 60.93889, "stage_name": "permutation_cycle", "status": "training", "step": 1526, "tokens_per_sec_per_gpu": 106084.264052, "vram_gb": 20.709735} {"adam_lr": 1.609201650659648e-05, "dt_ms": 158.345222, "global_tokens_per_sec": 103470.125237, "grad_norm": 36.5, "loss": 1.897536, "lr_stage": "cosine_decay", "mfu": 59.437228, "stage_name": "permutation_cycle", "status": "training", "step": 1527, "tokens_per_sec_per_gpu": 103470.125237, "vram_gb": 20.709735} {"adam_lr": 1.602793478692419e-05, "dt_ms": 158.520222, "global_tokens_per_sec": 103355.898845, "grad_norm": 23.0, "loss": 1.90475, "lr_stage": "cosine_decay", "mfu": 59.371612, "stage_name": "permutation_cycle", "status": "training", "step": 1528, "tokens_per_sec_per_gpu": 103355.898845, "vram_gb": 20.709735} {"adam_lr": 1.596395655205411e-05, "dt_ms": 155.91979, "global_tokens_per_sec": 105079.669308, "grad_norm": 19.875, "loss": 2.152834, "lr_stage": "cosine_decay", "mfu": 60.361812, "stage_name": "permutation_cycle", "status": "training", "step": 1529, "tokens_per_sec_per_gpu": 105079.669308, "vram_gb": 20.709735} {"adam_lr": 1.5900081996875083e-05, "dt_ms": 154.565811, "global_tokens_per_sec": 106000.155386, "grad_norm": 27.0, "loss": 2.432672, "lr_stage": "cosine_decay", "mfu": 60.890575, "stage_name": "permutation_cycle", "status": "training", "step": 1530, "tokens_per_sec_per_gpu": 106000.155386, "vram_gb": 20.709735} {"adam_lr": 1.583631131596005e-05, "dt_ms": 155.75552, "global_tokens_per_sec": 105190.493499, "grad_norm": 23.875, "loss": 1.831898, "lr_stage": "cosine_decay", "mfu": 60.425474, "stage_name": "permutation_cycle", "status": "training", "step": 1531, "tokens_per_sec_per_gpu": 105190.493499, "vram_gb": 20.709735} {"adam_lr": 1.5772644703565565e-05, "dt_ms": 155.445337, "global_tokens_per_sec": 105400.39531, "grad_norm": 31.75, "loss": 2.695769, "lr_stage": "cosine_decay", "mfu": 60.546049, "stage_name": "permutation_cycle", "status": "training", "step": 1532, "tokens_per_sec_per_gpu": 105400.39531, "vram_gb": 20.709735} {"adam_lr": 1.570908235363119e-05, "dt_ms": 155.856848, "global_tokens_per_sec": 105122.105542, "grad_norm": 30.125, "loss": 2.591353, "lr_stage": "cosine_decay", "mfu": 60.386189, "stage_name": "permutation_cycle", "status": "training", "step": 1533, "tokens_per_sec_per_gpu": 105122.105542, "vram_gb": 20.709735} {"adam_lr": 1.5645624459778856e-05, "dt_ms": 154.556751, "global_tokens_per_sec": 106006.36897, "grad_norm": 28.25, "loss": 2.014598, "lr_stage": "cosine_decay", "mfu": 60.894144, "stage_name": "permutation_cycle", "status": "training", "step": 1534, "tokens_per_sec_per_gpu": 106006.36897, "vram_gb": 20.709735} {"adam_lr": 1.5582271215312294e-05, "dt_ms": 155.417681, "global_tokens_per_sec": 105419.151296, "grad_norm": 29.625, "loss": 2.067754, "lr_stage": "cosine_decay", "mfu": 60.556824, "stage_name": "permutation_cycle", "status": "training", "step": 1535, "tokens_per_sec_per_gpu": 105419.151296, "vram_gb": 20.709735} {"adam_lr": 1.551902281321651e-05, "dt_ms": 154.292822, "global_tokens_per_sec": 106187.700762, "grad_norm": 26.875, "loss": 1.60017, "lr_stage": "cosine_decay", "mfu": 60.998308, "stage_name": "permutation_cycle", "status": "training", "step": 1536, "tokens_per_sec_per_gpu": 106187.700762, "vram_gb": 20.709735} {"adam_lr": 1.5455879446157083e-05, "dt_ms": 154.501915, "global_tokens_per_sec": 106043.993062, "grad_norm": 26.0, "loss": 1.767343, "lr_stage": "cosine_decay", "mfu": 60.915757, "stage_name": "permutation_cycle", "status": "training", "step": 1537, "tokens_per_sec_per_gpu": 106043.993062, "vram_gb": 20.709735} {"adam_lr": 1.5392841306479666e-05, "dt_ms": 155.346394, "global_tokens_per_sec": 105467.527259, "grad_norm": 21.875, "loss": 2.26185, "lr_stage": "cosine_decay", "mfu": 60.584613, "stage_name": "permutation_cycle", "status": "training", "step": 1538, "tokens_per_sec_per_gpu": 105467.527259, "vram_gb": 20.709735} {"adam_lr": 1.5329908586209347e-05, "dt_ms": 154.856443, "global_tokens_per_sec": 105801.215886, "grad_norm": 26.125, "loss": 1.934755, "lr_stage": "cosine_decay", "mfu": 60.776296, "stage_name": "permutation_cycle", "status": "training", "step": 1539, "tokens_per_sec_per_gpu": 105801.215886, "vram_gb": 20.709735} {"adam_lr": 1.526708147705013e-05, "dt_ms": 154.598236, "global_tokens_per_sec": 105977.923261, "grad_norm": 36.25, "loss": 2.802563, "lr_stage": "cosine_decay", "mfu": 60.877804, "stage_name": "permutation_cycle", "status": "training", "step": 1540, "tokens_per_sec_per_gpu": 105977.923261, "vram_gb": 20.709735} {"adam_lr": 1.5204360170384286e-05, "dt_ms": 156.587124, "global_tokens_per_sec": 104631.847083, "grad_norm": 30.375, "loss": 2.308371, "lr_stage": "cosine_decay", "mfu": 60.104566, "stage_name": "permutation_cycle", "status": "training", "step": 1541, "tokens_per_sec_per_gpu": 104631.847083, "vram_gb": 20.709735} {"adam_lr": 1.5141744857271778e-05, "dt_ms": 153.536558, "global_tokens_per_sec": 106710.741711, "grad_norm": 37.75, "loss": 2.352229, "lr_stage": "cosine_decay", "mfu": 61.298763, "stage_name": "permutation_cycle", "status": "training", "step": 1542, "tokens_per_sec_per_gpu": 106710.741711, "vram_gb": 20.709735} {"adam_lr": 1.5079235728449714e-05, "dt_ms": 155.757904, "global_tokens_per_sec": 105188.883348, "grad_norm": 36.75, "loss": 1.850581, "lr_stage": "cosine_decay", "mfu": 60.424549, "stage_name": "permutation_cycle", "status": "training", "step": 1543, "tokens_per_sec_per_gpu": 105188.883348, "vram_gb": 20.709735} {"adam_lr": 1.5016832974331724e-05, "dt_ms": 154.94442, "global_tokens_per_sec": 105741.142629, "grad_norm": 21.375, "loss": 2.770085, "lr_stage": "cosine_decay", "mfu": 60.741788, "stage_name": "permutation_cycle", "status": "training", "step": 1544, "tokens_per_sec_per_gpu": 105741.142629, "vram_gb": 20.709735} {"adam_lr": 1.4954536785007456e-05, "dt_ms": 155.301094, "global_tokens_per_sec": 105498.290915, "grad_norm": 25.375, "loss": 1.444714, "lr_stage": "cosine_decay", "mfu": 60.602284, "stage_name": "permutation_cycle", "status": "training", "step": 1545, "tokens_per_sec_per_gpu": 105498.290915, "vram_gb": 20.709735} {"adam_lr": 1.4892347350241881e-05, "dt_ms": 154.163599, "global_tokens_per_sec": 106276.709319, "grad_norm": 25.75, "loss": 2.058682, "lr_stage": "cosine_decay", "mfu": 61.049438, "stage_name": "permutation_cycle", "status": "training", "step": 1546, "tokens_per_sec_per_gpu": 106276.709319, "vram_gb": 20.709735} {"adam_lr": 1.4830264859474813e-05, "dt_ms": 154.961348, "global_tokens_per_sec": 105729.591642, "grad_norm": 27.75, "loss": 2.306021, "lr_stage": "cosine_decay", "mfu": 60.735153, "stage_name": "permutation_cycle", "status": "training", "step": 1547, "tokens_per_sec_per_gpu": 105729.591642, "vram_gb": 20.709735} {"adam_lr": 1.4768289501820265e-05, "dt_ms": 154.03223, "global_tokens_per_sec": 106367.348962, "grad_norm": 23.75, "loss": 1.79309, "lr_stage": "cosine_decay", "mfu": 61.101505, "stage_name": "permutation_cycle", "status": "training", "step": 1548, "tokens_per_sec_per_gpu": 106367.348962, "vram_gb": 20.709735} {"adam_lr": 1.470642146606595e-05, "dt_ms": 153.891563, "global_tokens_per_sec": 106464.575682, "grad_norm": 31.875, "loss": 1.930525, "lr_stage": "cosine_decay", "mfu": 61.157356, "stage_name": "permutation_cycle", "status": "training", "step": 1549, "tokens_per_sec_per_gpu": 106464.575682, "vram_gb": 20.709735} {"adam_lr": 1.4644660940672627e-05, "dt_ms": 154.911757, "global_tokens_per_sec": 105763.438286, "grad_norm": 30.25, "loss": 2.14329, "lr_stage": "cosine_decay", "mfu": 60.754595, "stage_name": "permutation_cycle", "status": "training", "step": 1550, "tokens_per_sec_per_gpu": 105763.438286, "vram_gb": 20.709735} {"adam_lr": 1.4583008113773566e-05, "dt_ms": 156.369686, "global_tokens_per_sec": 104777.341477, "grad_norm": 26.125, "loss": 2.088091, "lr_stage": "cosine_decay", "mfu": 60.188143, "stage_name": "permutation_cycle", "status": "training", "step": 1551, "tokens_per_sec_per_gpu": 104777.341477, "vram_gb": 20.709735} {"adam_lr": 1.4521463173173965e-05, "dt_ms": 154.50263, "global_tokens_per_sec": 106043.502141, "grad_norm": 29.375, "loss": 2.478299, "lr_stage": "cosine_decay", "mfu": 60.915475, "stage_name": "permutation_cycle", "status": "training", "step": 1552, "tokens_per_sec_per_gpu": 106043.502141, "vram_gb": 20.709735} {"adam_lr": 1.4460026306350377e-05, "dt_ms": 154.631853, "global_tokens_per_sec": 105954.883623, "grad_norm": 32.25, "loss": 1.843433, "lr_stage": "cosine_decay", "mfu": 60.864569, "stage_name": "permutation_cycle", "status": "training", "step": 1553, "tokens_per_sec_per_gpu": 105954.883623, "vram_gb": 20.709735} {"adam_lr": 1.439869770045018e-05, "dt_ms": 155.545473, "global_tokens_per_sec": 105332.541498, "grad_norm": 32.25, "loss": 2.661978, "lr_stage": "cosine_decay", "mfu": 60.507072, "stage_name": "permutation_cycle", "status": "training", "step": 1554, "tokens_per_sec_per_gpu": 105332.541498, "vram_gb": 20.709735} {"adam_lr": 1.4337477542290928e-05, "dt_ms": 155.905962, "global_tokens_per_sec": 105088.989483, "grad_norm": 21.375, "loss": 2.072887, "lr_stage": "cosine_decay", "mfu": 60.367166, "stage_name": "permutation_cycle", "status": "training", "step": 1555, "tokens_per_sec_per_gpu": 105088.989483, "vram_gb": 20.709735} {"adam_lr": 1.4276366018359844e-05, "dt_ms": 155.311108, "global_tokens_per_sec": 105491.488983, "grad_norm": 30.375, "loss": 2.394375, "lr_stage": "cosine_decay", "mfu": 60.598377, "stage_name": "permutation_cycle", "status": "training", "step": 1556, "tokens_per_sec_per_gpu": 105491.488983, "vram_gb": 20.709735} {"adam_lr": 1.4215363314813207e-05, "dt_ms": 156.137228, "global_tokens_per_sec": 104933.334661, "grad_norm": 31.625, "loss": 2.14662, "lr_stage": "cosine_decay", "mfu": 60.277752, "stage_name": "permutation_cycle", "status": "training", "step": 1557, "tokens_per_sec_per_gpu": 104933.334661, "vram_gb": 20.709735} {"adam_lr": 1.4154469617475863e-05, "dt_ms": 154.753208, "global_tokens_per_sec": 105871.795453, "grad_norm": 36.75, "loss": 2.307904, "lr_stage": "cosine_decay", "mfu": 60.81684, "stage_name": "permutation_cycle", "status": "training", "step": 1558, "tokens_per_sec_per_gpu": 105871.795453, "vram_gb": 20.709735} {"adam_lr": 1.4093685111840566e-05, "dt_ms": 154.395342, "global_tokens_per_sec": 106117.19111, "grad_norm": 32.5, "loss": 1.522845, "lr_stage": "cosine_decay", "mfu": 60.957805, "stage_name": "permutation_cycle", "status": "training", "step": 1559, "tokens_per_sec_per_gpu": 106117.19111, "vram_gb": 20.709735} {"adam_lr": 1.4033009983067452e-05, "dt_ms": 156.194448, "global_tokens_per_sec": 104894.893259, "grad_norm": 25.5, "loss": 2.748383, "lr_stage": "cosine_decay", "mfu": 60.25567, "stage_name": "permutation_cycle", "status": "training", "step": 1560, "tokens_per_sec_per_gpu": 104894.893259, "vram_gb": 20.709735} {"adam_lr": 1.3972444415983493e-05, "dt_ms": 156.027794, "global_tokens_per_sec": 105006.932368, "grad_norm": 30.125, "loss": 2.905304, "lr_stage": "cosine_decay", "mfu": 60.320029, "stage_name": "permutation_cycle", "status": "training", "step": 1561, "tokens_per_sec_per_gpu": 105006.932368, "vram_gb": 20.709735} {"adam_lr": 1.3911988595081893e-05, "dt_ms": 152.576208, "global_tokens_per_sec": 107382.403865, "grad_norm": 30.5, "loss": 1.52532, "lr_stage": "cosine_decay", "mfu": 61.684592, "stage_name": "permutation_cycle", "status": "training", "step": 1562, "tokens_per_sec_per_gpu": 107382.403865, "vram_gb": 20.709735} {"adam_lr": 1.3851642704521595e-05, "dt_ms": 156.000614, "global_tokens_per_sec": 105025.227545, "grad_norm": 29.875, "loss": 2.468709, "lr_stage": "cosine_decay", "mfu": 60.330539, "stage_name": "permutation_cycle", "status": "training", "step": 1563, "tokens_per_sec_per_gpu": 105025.227545, "vram_gb": 20.709735} {"adam_lr": 1.3791406928126638e-05, "dt_ms": 154.985905, "global_tokens_per_sec": 105712.839064, "grad_norm": 30.625, "loss": 2.143734, "lr_stage": "cosine_decay", "mfu": 60.725529, "stage_name": "permutation_cycle", "status": "training", "step": 1564, "tokens_per_sec_per_gpu": 105712.839064, "vram_gb": 20.709735} {"adam_lr": 1.373128144938563e-05, "dt_ms": 155.136347, "global_tokens_per_sec": 105610.324957, "grad_norm": 26.125, "loss": 1.964124, "lr_stage": "cosine_decay", "mfu": 60.666641, "stage_name": "permutation_cycle", "status": "training", "step": 1565, "tokens_per_sec_per_gpu": 105610.324957, "vram_gb": 20.709735} {"adam_lr": 1.367126645145121e-05, "dt_ms": 156.233311, "global_tokens_per_sec": 104868.801196, "grad_norm": 27.0, "loss": 1.814708, "lr_stage": "cosine_decay", "mfu": 60.240681, "stage_name": "permutation_cycle", "status": "training", "step": 1566, "tokens_per_sec_per_gpu": 104868.801196, "vram_gb": 20.709735} {"adam_lr": 1.361136211713948e-05, "dt_ms": 154.504776, "global_tokens_per_sec": 106042.029406, "grad_norm": 27.625, "loss": 2.460792, "lr_stage": "cosine_decay", "mfu": 60.914629, "stage_name": "permutation_cycle", "status": "training", "step": 1567, "tokens_per_sec_per_gpu": 106042.029406, "vram_gb": 20.709735} {"adam_lr": 1.3551568628929434e-05, "dt_ms": 157.525063, "global_tokens_per_sec": 104008.846171, "grad_norm": 27.25, "loss": 2.553762, "lr_stage": "cosine_decay", "mfu": 59.74669, "stage_name": "permutation_cycle", "status": "training", "step": 1568, "tokens_per_sec_per_gpu": 104008.846171, "vram_gb": 20.709735} {"adam_lr": 1.349188616896238e-05, "dt_ms": 157.85265, "global_tokens_per_sec": 103792.999562, "grad_norm": 23.25, "loss": 2.707357, "lr_stage": "cosine_decay", "mfu": 59.622699, "stage_name": "permutation_cycle", "status": "training", "step": 1569, "tokens_per_sec_per_gpu": 103792.999562, "vram_gb": 20.709735} {"adam_lr": 1.3432314919041478e-05, "dt_ms": 157.191753, "global_tokens_per_sec": 104229.386383, "grad_norm": 34.0, "loss": 1.870573, "lr_stage": "cosine_decay", "mfu": 59.873377, "stage_name": "permutation_cycle", "status": "training", "step": 1570, "tokens_per_sec_per_gpu": 104229.386383, "vram_gb": 20.709735} {"adam_lr": 1.3372855060631068e-05, "dt_ms": 156.836033, "global_tokens_per_sec": 104465.789528, "grad_norm": 31.0, "loss": 2.150841, "lr_stage": "cosine_decay", "mfu": 60.009176, "stage_name": "permutation_cycle", "status": "training", "step": 1571, "tokens_per_sec_per_gpu": 104465.789528, "vram_gb": 20.709735} {"adam_lr": 1.3313506774856177e-05, "dt_ms": 155.469179, "global_tokens_per_sec": 105384.231712, "grad_norm": 30.375, "loss": 2.133451, "lr_stage": "cosine_decay", "mfu": 60.536764, "stage_name": "permutation_cycle", "status": "training", "step": 1572, "tokens_per_sec_per_gpu": 105384.231712, "vram_gb": 20.709735} {"adam_lr": 1.3254270242502003e-05, "dt_ms": 155.828714, "global_tokens_per_sec": 105141.084338, "grad_norm": 31.5, "loss": 1.947017, "lr_stage": "cosine_decay", "mfu": 60.397091, "stage_name": "permutation_cycle", "status": "training", "step": 1573, "tokens_per_sec_per_gpu": 105141.084338, "vram_gb": 20.709735} {"adam_lr": 1.3195145644013285e-05, "dt_ms": 153.589725, "global_tokens_per_sec": 106673.802217, "grad_norm": 29.5, "loss": 2.051452, "lr_stage": "cosine_decay", "mfu": 61.277543, "stage_name": "permutation_cycle", "status": "training", "step": 1574, "tokens_per_sec_per_gpu": 106673.802217, "vram_gb": 20.709735} {"adam_lr": 1.3136133159493802e-05, "dt_ms": 155.866146, "global_tokens_per_sec": 105115.834395, "grad_norm": 27.0, "loss": 2.238137, "lr_stage": "cosine_decay", "mfu": 60.382587, "stage_name": "permutation_cycle", "status": "training", "step": 1575, "tokens_per_sec_per_gpu": 105115.834395, "vram_gb": 20.709735} {"adam_lr": 1.3077232968705805e-05, "dt_ms": 156.075001, "global_tokens_per_sec": 104975.17168, "grad_norm": 20.625, "loss": 2.294025, "lr_stage": "cosine_decay", "mfu": 60.301785, "stage_name": "permutation_cycle", "status": "training", "step": 1576, "tokens_per_sec_per_gpu": 104975.17168, "vram_gb": 20.709735} {"adam_lr": 1.301844525106951e-05, "dt_ms": 154.225349, "global_tokens_per_sec": 106234.157102, "grad_norm": 26.875, "loss": 1.475673, "lr_stage": "cosine_decay", "mfu": 61.024994, "stage_name": "permutation_cycle", "status": "training", "step": 1577, "tokens_per_sec_per_gpu": 106234.157102, "vram_gb": 20.709735} {"adam_lr": 1.29597701856625e-05, "dt_ms": 154.862404, "global_tokens_per_sec": 105797.143726, "grad_norm": 31.75, "loss": 2.407896, "lr_stage": "cosine_decay", "mfu": 60.773957, "stage_name": "permutation_cycle", "status": "training", "step": 1578, "tokens_per_sec_per_gpu": 105797.143726, "vram_gb": 20.709735} {"adam_lr": 1.2901207951219186e-05, "dt_ms": 155.794859, "global_tokens_per_sec": 105163.932316, "grad_norm": 29.125, "loss": 2.64734, "lr_stage": "cosine_decay", "mfu": 60.410216, "stage_name": "permutation_cycle", "status": "training", "step": 1579, "tokens_per_sec_per_gpu": 105163.932316, "vram_gb": 20.709735} {"adam_lr": 1.2842758726130283e-05, "dt_ms": 156.77619, "global_tokens_per_sec": 104505.665181, "grad_norm": 30.125, "loss": 2.949432, "lr_stage": "cosine_decay", "mfu": 60.032082, "stage_name": "permutation_cycle", "status": "training", "step": 1580, "tokens_per_sec_per_gpu": 104505.665181, "vram_gb": 20.709735} {"adam_lr": 1.2784422688442294e-05, "dt_ms": 159.643173, "global_tokens_per_sec": 102628.87958, "grad_norm": 39.75, "loss": 3.583999, "lr_stage": "cosine_decay", "mfu": 58.953984, "stage_name": "permutation_cycle", "status": "training", "step": 1581, "tokens_per_sec_per_gpu": 102628.87958, "vram_gb": 20.709735} {"adam_lr": 1.2726200015856892e-05, "dt_ms": 159.331322, "global_tokens_per_sec": 102829.75013, "grad_norm": 29.75, "loss": 2.564253, "lr_stage": "cosine_decay", "mfu": 59.069372, "stage_name": "permutation_cycle", "status": "training", "step": 1582, "tokens_per_sec_per_gpu": 102829.75013, "vram_gb": 20.709735} {"adam_lr": 1.2668090885730438e-05, "dt_ms": 159.829855, "global_tokens_per_sec": 102509.008743, "grad_norm": 30.75, "loss": 3.271471, "lr_stage": "cosine_decay", "mfu": 58.885126, "stage_name": "permutation_cycle", "status": "training", "step": 1583, "tokens_per_sec_per_gpu": 102509.008743, "vram_gb": 20.709735} {"adam_lr": 1.2610095475073414e-05, "dt_ms": 154.965639, "global_tokens_per_sec": 105726.663624, "grad_norm": 19.375, "loss": 2.133569, "lr_stage": "cosine_decay", "mfu": 60.733471, "stage_name": "permutation_cycle", "status": "training", "step": 1584, "tokens_per_sec_per_gpu": 105726.663624, "vram_gb": 20.709735} {"adam_lr": 1.2552213960549891e-05, "dt_ms": 154.712439, "global_tokens_per_sec": 105899.694621, "grad_norm": 31.625, "loss": 1.62525, "lr_stage": "cosine_decay", "mfu": 60.832866, "stage_name": "permutation_cycle", "status": "training", "step": 1585, "tokens_per_sec_per_gpu": 105899.694621, "vram_gb": 20.709735} {"adam_lr": 1.2494446518477022e-05, "dt_ms": 155.071974, "global_tokens_per_sec": 105654.165601, "grad_norm": 31.5, "loss": 2.872715, "lr_stage": "cosine_decay", "mfu": 60.691825, "stage_name": "permutation_cycle", "status": "training", "step": 1586, "tokens_per_sec_per_gpu": 105654.165601, "vram_gb": 20.709735} {"adam_lr": 1.2436793324824447e-05, "dt_ms": 154.832125, "global_tokens_per_sec": 105817.833545, "grad_norm": 38.0, "loss": 1.31335, "lr_stage": "cosine_decay", "mfu": 60.785842, "stage_name": "permutation_cycle", "status": "training", "step": 1587, "tokens_per_sec_per_gpu": 105817.833545, "vram_gb": 20.709735} {"adam_lr": 1.2379254555213788e-05, "dt_ms": 153.33271, "global_tokens_per_sec": 106852.608107, "grad_norm": 16.25, "loss": 2.028746, "lr_stage": "cosine_decay", "mfu": 61.380256, "stage_name": "permutation_cycle", "status": "training", "step": 1588, "tokens_per_sec_per_gpu": 106852.608107, "vram_gb": 20.709735} {"adam_lr": 1.2321830384918116e-05, "dt_ms": 154.701233, "global_tokens_per_sec": 105907.365389, "grad_norm": 27.125, "loss": 2.739851, "lr_stage": "cosine_decay", "mfu": 60.837273, "stage_name": "permutation_cycle", "status": "training", "step": 1589, "tokens_per_sec_per_gpu": 105907.365389, "vram_gb": 20.709735} {"adam_lr": 1.22645209888614e-05, "dt_ms": 157.42135, "global_tokens_per_sec": 104077.369112, "grad_norm": 27.5, "loss": 2.678343, "lr_stage": "cosine_decay", "mfu": 59.786052, "stage_name": "permutation_cycle", "status": "training", "step": 1590, "tokens_per_sec_per_gpu": 104077.369112, "vram_gb": 20.709735} {"adam_lr": 1.2207326541618025e-05, "dt_ms": 155.323744, "global_tokens_per_sec": 105482.906844, "grad_norm": 26.125, "loss": 1.752521, "lr_stage": "cosine_decay", "mfu": 60.593447, "stage_name": "permutation_cycle", "status": "training", "step": 1591, "tokens_per_sec_per_gpu": 105482.906844, "vram_gb": 20.709735} {"adam_lr": 1.2150247217412186e-05, "dt_ms": 155.55644, "global_tokens_per_sec": 105325.115198, "grad_norm": 27.5, "loss": 2.591386, "lr_stage": "cosine_decay", "mfu": 60.502806, "stage_name": "permutation_cycle", "status": "training", "step": 1592, "tokens_per_sec_per_gpu": 105325.115198, "vram_gb": 20.709735} {"adam_lr": 1.2093283190117399e-05, "dt_ms": 155.030489, "global_tokens_per_sec": 105682.437752, "grad_norm": 29.875, "loss": 1.978847, "lr_stage": "cosine_decay", "mfu": 60.708065, "stage_name": "permutation_cycle", "status": "training", "step": 1593, "tokens_per_sec_per_gpu": 105682.437752, "vram_gb": 20.709735} {"adam_lr": 1.203643463325596e-05, "dt_ms": 154.564142, "global_tokens_per_sec": 106001.299939, "grad_norm": 29.75, "loss": 1.898116, "lr_stage": "cosine_decay", "mfu": 60.891232, "stage_name": "permutation_cycle", "status": "training", "step": 1594, "tokens_per_sec_per_gpu": 106001.299939, "vram_gb": 20.709735} {"adam_lr": 1.1979701719998453e-05, "dt_ms": 158.022404, "global_tokens_per_sec": 103681.500943, "grad_norm": 26.0, "loss": 2.829123, "lr_stage": "cosine_decay", "mfu": 59.55865, "stage_name": "permutation_cycle", "status": "training", "step": 1595, "tokens_per_sec_per_gpu": 103681.500943, "vram_gb": 20.709735} {"adam_lr": 1.1923084623163172e-05, "dt_ms": 157.4049, "global_tokens_per_sec": 104088.246566, "grad_norm": 25.25, "loss": 2.277301, "lr_stage": "cosine_decay", "mfu": 59.792301, "stage_name": "permutation_cycle", "status": "training", "step": 1596, "tokens_per_sec_per_gpu": 104088.246566, "vram_gb": 20.709735} {"adam_lr": 1.1866583515215596e-05, "dt_ms": 154.893875, "global_tokens_per_sec": 105775.647921, "grad_norm": 19.375, "loss": 2.296304, "lr_stage": "cosine_decay", "mfu": 60.761609, "stage_name": "permutation_cycle", "status": "training", "step": 1597, "tokens_per_sec_per_gpu": 105775.647921, "vram_gb": 20.709735} {"adam_lr": 1.1810198568267905e-05, "dt_ms": 155.586481, "global_tokens_per_sec": 105304.778955, "grad_norm": 40.0, "loss": 2.758254, "lr_stage": "cosine_decay", "mfu": 60.491124, "stage_name": "permutation_cycle", "status": "training", "step": 1598, "tokens_per_sec_per_gpu": 105304.778955, "vram_gb": 20.709735} {"adam_lr": 1.1753929954078413e-05, "dt_ms": 157.236338, "global_tokens_per_sec": 104199.832199, "grad_norm": 35.75, "loss": 2.46994, "lr_stage": "cosine_decay", "mfu": 59.8564, "stage_name": "permutation_cycle", "status": "training", "step": 1599, "tokens_per_sec_per_gpu": 104199.832199, "vram_gb": 20.709735} {"adam_lr": 1.1697777844051105e-05, "dt_ms": 153.292894, "global_tokens_per_sec": 106880.361729, "grad_norm": 41.75, "loss": 1.502969, "lr_stage": "cosine_decay", "mfu": 61.396199, "stage_name": "permutation_cycle", "status": "training", "step": 1600, "tokens_per_sec_per_gpu": 106880.361729, "vram_gb": 20.709735} {"adam_lr": 1.164174240923503e-05, "dt_ms": 155.72691, "global_tokens_per_sec": 105209.819152, "grad_norm": 24.625, "loss": 2.79016, "lr_stage": "cosine_decay", "mfu": 60.436575, "stage_name": "permutation_cycle", "status": "training", "step": 1601, "tokens_per_sec_per_gpu": 105209.819152, "vram_gb": 20.709735} {"adam_lr": 1.1585823820323843e-05, "dt_ms": 155.386209, "global_tokens_per_sec": 105440.502436, "grad_norm": 25.25, "loss": 1.744532, "lr_stage": "cosine_decay", "mfu": 60.569089, "stage_name": "permutation_cycle", "status": "training", "step": 1602, "tokens_per_sec_per_gpu": 105440.502436, "vram_gb": 20.709735} {"adam_lr": 1.1530022247655252e-05, "dt_ms": 154.67453, "global_tokens_per_sec": 105925.649148, "grad_norm": 33.5, "loss": 2.013561, "lr_stage": "cosine_decay", "mfu": 60.847775, "stage_name": "permutation_cycle", "status": "training", "step": 1603, "tokens_per_sec_per_gpu": 105925.649148, "vram_gb": 20.709735} {"adam_lr": 1.1474337861210543e-05, "dt_ms": 155.201674, "global_tokens_per_sec": 105565.872002, "grad_norm": 21.625, "loss": 2.223469, "lr_stage": "cosine_decay", "mfu": 60.641106, "stage_name": "permutation_cycle", "status": "training", "step": 1604, "tokens_per_sec_per_gpu": 105565.872002, "vram_gb": 20.709735} {"adam_lr": 1.1418770830614013e-05, "dt_ms": 154.57058, "global_tokens_per_sec": 105996.885371, "grad_norm": 19.125, "loss": 1.558846, "lr_stage": "cosine_decay", "mfu": 60.888696, "stage_name": "permutation_cycle", "status": "training", "step": 1605, "tokens_per_sec_per_gpu": 105996.885371, "vram_gb": 20.709735} {"adam_lr": 1.1363321325132447e-05, "dt_ms": 155.040979, "global_tokens_per_sec": 105675.287043, "grad_norm": 24.5, "loss": 2.00663, "lr_stage": "cosine_decay", "mfu": 60.703958, "stage_name": "permutation_cycle", "status": "training", "step": 1606, "tokens_per_sec_per_gpu": 105675.287043, "vram_gb": 20.709735} {"adam_lr": 1.1307989513674693e-05, "dt_ms": 156.528711, "global_tokens_per_sec": 104670.893039, "grad_norm": 15.875, "loss": 2.23631, "lr_stage": "cosine_decay", "mfu": 60.126995, "stage_name": "permutation_cycle", "status": "training", "step": 1607, "tokens_per_sec_per_gpu": 104670.893039, "vram_gb": 20.709735} {"adam_lr": 1.1252775564791024e-05, "dt_ms": 155.325174, "global_tokens_per_sec": 105481.935369, "grad_norm": 21.875, "loss": 1.948438, "lr_stage": "cosine_decay", "mfu": 60.592889, "stage_name": "permutation_cycle", "status": "training", "step": 1608, "tokens_per_sec_per_gpu": 105481.935369, "vram_gb": 20.709735} {"adam_lr": 1.1197679646672698e-05, "dt_ms": 156.159163, "global_tokens_per_sec": 104918.59546, "grad_norm": 26.5, "loss": 2.080355, "lr_stage": "cosine_decay", "mfu": 60.269285, "stage_name": "permutation_cycle", "status": "training", "step": 1609, "tokens_per_sec_per_gpu": 104918.59546, "vram_gb": 20.709735} {"adam_lr": 1.1142701927151456e-05, "dt_ms": 156.641245, "global_tokens_per_sec": 104595.6958, "grad_norm": 20.5, "loss": 2.102447, "lr_stage": "cosine_decay", "mfu": 60.083799, "stage_name": "permutation_cycle", "status": "training", "step": 1610, "tokens_per_sec_per_gpu": 104595.6958, "vram_gb": 20.709735} {"adam_lr": 1.1087842573698953e-05, "dt_ms": 155.07412, "global_tokens_per_sec": 105652.70366, "grad_norm": 37.0, "loss": 2.564031, "lr_stage": "cosine_decay", "mfu": 60.690985, "stage_name": "permutation_cycle", "status": "training", "step": 1611, "tokens_per_sec_per_gpu": 105652.70366, "vram_gb": 20.709735} {"adam_lr": 1.1033101753426283e-05, "dt_ms": 158.018351, "global_tokens_per_sec": 103684.160337, "grad_norm": 28.875, "loss": 2.343783, "lr_stage": "cosine_decay", "mfu": 59.560178, "stage_name": "permutation_cycle", "status": "training", "step": 1612, "tokens_per_sec_per_gpu": 103684.160337, "vram_gb": 20.709735} {"adam_lr": 1.097847963308351e-05, "dt_ms": 158.984661, "global_tokens_per_sec": 103053.966882, "grad_norm": 20.75, "loss": 2.527651, "lr_stage": "cosine_decay", "mfu": 59.198171, "stage_name": "permutation_cycle", "status": "training", "step": 1613, "tokens_per_sec_per_gpu": 103053.966882, "vram_gb": 20.709735} {"adam_lr": 1.0923976379059058e-05, "dt_ms": 158.776045, "global_tokens_per_sec": 103189.369756, "grad_norm": 20.5, "loss": 2.578746, "lr_stage": "cosine_decay", "mfu": 59.275951, "stage_name": "permutation_cycle", "status": "training", "step": 1614, "tokens_per_sec_per_gpu": 103189.369756, "vram_gb": 20.709735} {"adam_lr": 1.0869592157379304e-05, "dt_ms": 159.228325, "global_tokens_per_sec": 102896.265544, "grad_norm": 31.875, "loss": 1.867867, "lr_stage": "cosine_decay", "mfu": 59.107581, "stage_name": "permutation_cycle", "status": "training", "step": 1615, "tokens_per_sec_per_gpu": 102896.265544, "vram_gb": 20.709735} {"adam_lr": 1.0815327133708015e-05, "dt_ms": 153.603315, "global_tokens_per_sec": 106664.364388, "grad_norm": 30.75, "loss": 1.620594, "lr_stage": "cosine_decay", "mfu": 61.272122, "stage_name": "permutation_cycle", "status": "training", "step": 1616, "tokens_per_sec_per_gpu": 106664.364388, "vram_gb": 20.709735} {"adam_lr": 1.0761181473345849e-05, "dt_ms": 155.006886, "global_tokens_per_sec": 105698.530385, "grad_norm": 33.25, "loss": 2.685536, "lr_stage": "cosine_decay", "mfu": 60.71731, "stage_name": "permutation_cycle", "status": "training", "step": 1617, "tokens_per_sec_per_gpu": 105698.530385, "vram_gb": 20.709735} {"adam_lr": 1.0707155341229901e-05, "dt_ms": 155.421734, "global_tokens_per_sec": 105416.402156, "grad_norm": 41.0, "loss": 2.742507, "lr_stage": "cosine_decay", "mfu": 60.555244, "stage_name": "permutation_cycle", "status": "training", "step": 1618, "tokens_per_sec_per_gpu": 105416.402156, "vram_gb": 20.709735} {"adam_lr": 1.0653248901933138e-05, "dt_ms": 155.940056, "global_tokens_per_sec": 105066.013418, "grad_norm": 17.375, "loss": 2.098353, "lr_stage": "cosine_decay", "mfu": 60.353968, "stage_name": "permutation_cycle", "status": "training", "step": 1619, "tokens_per_sec_per_gpu": 105066.013418, "vram_gb": 20.709735} {"adam_lr": 1.0599462319663905e-05, "dt_ms": 154.761791, "global_tokens_per_sec": 105865.923817, "grad_norm": 31.625, "loss": 1.581123, "lr_stage": "cosine_decay", "mfu": 60.813467, "stage_name": "permutation_cycle", "status": "training", "step": 1620, "tokens_per_sec_per_gpu": 105865.923817, "vram_gb": 20.709735} {"adam_lr": 1.0545795758265476e-05, "dt_ms": 155.175924, "global_tokens_per_sec": 105583.389136, "grad_norm": 18.75, "loss": 1.358465, "lr_stage": "cosine_decay", "mfu": 60.651168, "stage_name": "permutation_cycle", "status": "training", "step": 1621, "tokens_per_sec_per_gpu": 105583.389136, "vram_gb": 20.709735} {"adam_lr": 1.049224938121548e-05, "dt_ms": 155.738115, "global_tokens_per_sec": 105202.249092, "grad_norm": 24.625, "loss": 2.144895, "lr_stage": "cosine_decay", "mfu": 60.432227, "stage_name": "permutation_cycle", "status": "training", "step": 1622, "tokens_per_sec_per_gpu": 105202.249092, "vram_gb": 20.709735} {"adam_lr": 1.0438823351625498e-05, "dt_ms": 158.770561, "global_tokens_per_sec": 103192.933717, "grad_norm": 28.5, "loss": 2.451661, "lr_stage": "cosine_decay", "mfu": 59.277998, "stage_name": "permutation_cycle", "status": "training", "step": 1623, "tokens_per_sec_per_gpu": 103192.933717, "vram_gb": 20.709735} {"adam_lr": 1.0385517832240471e-05, "dt_ms": 157.973528, "global_tokens_per_sec": 103713.579211, "grad_norm": 17.625, "loss": 2.081149, "lr_stage": "cosine_decay", "mfu": 59.577077, "stage_name": "permutation_cycle", "status": "training", "step": 1624, "tokens_per_sec_per_gpu": 103713.579211, "vram_gb": 20.709735} {"adam_lr": 1.0332332985438248e-05, "dt_ms": 155.240297, "global_tokens_per_sec": 105539.607197, "grad_norm": 32.5, "loss": 2.297833, "lr_stage": "cosine_decay", "mfu": 60.626018, "stage_name": "permutation_cycle", "status": "training", "step": 1625, "tokens_per_sec_per_gpu": 105539.607197, "vram_gb": 20.709735} {"adam_lr": 1.0279268973229089e-05, "dt_ms": 155.234814, "global_tokens_per_sec": 105543.335355, "grad_norm": 37.25, "loss": 3.135667, "lr_stage": "cosine_decay", "mfu": 60.62816, "stage_name": "permutation_cycle", "status": "training", "step": 1626, "tokens_per_sec_per_gpu": 105543.335355, "vram_gb": 20.709735} {"adam_lr": 1.0226325957255206e-05, "dt_ms": 155.767441, "global_tokens_per_sec": 105182.443239, "grad_norm": 31.5, "loss": 3.121526, "lr_stage": "cosine_decay", "mfu": 60.420849, "stage_name": "permutation_cycle", "status": "training", "step": 1627, "tokens_per_sec_per_gpu": 105182.443239, "vram_gb": 20.709735} {"adam_lr": 1.0173504098790187e-05, "dt_ms": 154.633999, "global_tokens_per_sec": 105953.413348, "grad_norm": 30.0, "loss": 2.346937, "lr_stage": "cosine_decay", "mfu": 60.863724, "stage_name": "permutation_cycle", "status": "training", "step": 1628, "tokens_per_sec_per_gpu": 105953.413348, "vram_gb": 20.709735} {"adam_lr": 1.0120803558738585e-05, "dt_ms": 154.872179, "global_tokens_per_sec": 105790.466064, "grad_norm": 33.25, "loss": 2.513784, "lr_stage": "cosine_decay", "mfu": 60.770121, "stage_name": "permutation_cycle", "status": "training", "step": 1629, "tokens_per_sec_per_gpu": 105790.466064, "vram_gb": 20.709735} {"adam_lr": 1.006822449763537e-05, "dt_ms": 158.895254, "global_tokens_per_sec": 103111.953149, "grad_norm": 22.125, "loss": 2.474931, "lr_stage": "cosine_decay", "mfu": 59.23148, "stage_name": "permutation_cycle", "status": "training", "step": 1630, "tokens_per_sec_per_gpu": 103111.953149, "vram_gb": 20.709735} {"adam_lr": 1.0015767075645471e-05, "dt_ms": 158.812523, "global_tokens_per_sec": 103165.667934, "grad_norm": 23.0, "loss": 1.993032, "lr_stage": "cosine_decay", "mfu": 59.262336, "stage_name": "permutation_cycle", "status": "training", "step": 1631, "tokens_per_sec_per_gpu": 103165.667934, "vram_gb": 20.709735} {"adam_lr": 9.963431452563332e-06, "dt_ms": 159.977436, "global_tokens_per_sec": 102414.442955, "grad_norm": 31.875, "loss": 2.724316, "lr_stage": "cosine_decay", "mfu": 58.830803, "stage_name": "permutation_cycle", "status": "training", "step": 1632, "tokens_per_sec_per_gpu": 102414.442955, "vram_gb": 20.709735} {"adam_lr": 9.911217787812304e-06, "dt_ms": 154.307842, "global_tokens_per_sec": 106177.36442, "grad_norm": 36.5, "loss": 1.987976, "lr_stage": "cosine_decay", "mfu": 60.99237, "stage_name": "permutation_cycle", "status": "training", "step": 1633, "tokens_per_sec_per_gpu": 106177.36442, "vram_gb": 20.709735} {"adam_lr": 9.859126240444283e-06, "dt_ms": 154.441357, "global_tokens_per_sec": 106085.574191, "grad_norm": 29.875, "loss": 1.813979, "lr_stage": "cosine_decay", "mfu": 60.939643, "stage_name": "permutation_cycle", "status": "training", "step": 1634, "tokens_per_sec_per_gpu": 106085.574191, "vram_gb": 20.709735} {"adam_lr": 9.807156969139136e-06, "dt_ms": 155.794859, "global_tokens_per_sec": 105163.932316, "grad_norm": 24.5, "loss": 2.336683, "lr_stage": "cosine_decay", "mfu": 60.410216, "stage_name": "permutation_cycle", "status": "training", "step": 1635, "tokens_per_sec_per_gpu": 105163.932316, "vram_gb": 20.709735} {"adam_lr": 9.755310132204298e-06, "dt_ms": 155.321598, "global_tokens_per_sec": 105484.36409, "grad_norm": 30.75, "loss": 1.87219, "lr_stage": "cosine_decay", "mfu": 60.594284, "stage_name": "permutation_cycle", "status": "training", "step": 1636, "tokens_per_sec_per_gpu": 105484.36409, "vram_gb": 20.709735} {"adam_lr": 9.703585887574218e-06, "dt_ms": 154.834032, "global_tokens_per_sec": 105816.530011, "grad_norm": 47.5, "loss": 3.038322, "lr_stage": "cosine_decay", "mfu": 60.785093, "stage_name": "permutation_cycle", "status": "training", "step": 1637, "tokens_per_sec_per_gpu": 105816.530011, "vram_gb": 20.709735} {"adam_lr": 9.651984392809914e-06, "dt_ms": 155.269861, "global_tokens_per_sec": 105519.512101, "grad_norm": 38.5, "loss": 3.361219, "lr_stage": "cosine_decay", "mfu": 60.614475, "stage_name": "permutation_cycle", "status": "training", "step": 1638, "tokens_per_sec_per_gpu": 105519.512101, "vram_gb": 20.709735} {"adam_lr": 9.600505805098487e-06, "dt_ms": 155.245781, "global_tokens_per_sec": 105535.879302, "grad_norm": 43.75, "loss": 2.188178, "lr_stage": "cosine_decay", "mfu": 60.623877, "stage_name": "permutation_cycle", "status": "training", "step": 1639, "tokens_per_sec_per_gpu": 105535.879302, "vram_gb": 20.709735} {"adam_lr": 9.549150281252633e-06, "dt_ms": 154.108047, "global_tokens_per_sec": 106315.019023, "grad_norm": 27.0, "loss": 2.026596, "lr_stage": "cosine_decay", "mfu": 61.071445, "stage_name": "permutation_cycle", "status": "training", "step": 1640, "tokens_per_sec_per_gpu": 106315.019023, "vram_gb": 20.709735} {"adam_lr": 9.497917977710208e-06, "dt_ms": 159.208775, "global_tokens_per_sec": 102908.900873, "grad_norm": 25.75, "loss": 1.878725, "lr_stage": "cosine_decay", "mfu": 59.114839, "stage_name": "permutation_cycle", "status": "training", "step": 1641, "tokens_per_sec_per_gpu": 102908.900873, "vram_gb": 20.709735} {"adam_lr": 9.446809050533678e-06, "dt_ms": 157.643795, "global_tokens_per_sec": 103930.509911, "grad_norm": 33.5, "loss": 2.386668, "lr_stage": "cosine_decay", "mfu": 59.701691, "stage_name": "permutation_cycle", "status": "training", "step": 1642, "tokens_per_sec_per_gpu": 103930.509911, "vram_gb": 20.709735} {"adam_lr": 9.395823655409685e-06, "dt_ms": 154.712439, "global_tokens_per_sec": 105899.694621, "grad_norm": 29.375, "loss": 2.885805, "lr_stage": "cosine_decay", "mfu": 60.832866, "stage_name": "permutation_cycle", "status": "training", "step": 1643, "tokens_per_sec_per_gpu": 105899.694621, "vram_gb": 20.709735} {"adam_lr": 9.344961947648623e-06, "dt_ms": 156.09622, "global_tokens_per_sec": 104960.901669, "grad_norm": 29.0, "loss": 2.565695, "lr_stage": "cosine_decay", "mfu": 60.293587, "stage_name": "permutation_cycle", "status": "training", "step": 1644, "tokens_per_sec_per_gpu": 104960.901669, "vram_gb": 20.709735} {"adam_lr": 9.294224082184045e-06, "dt_ms": 153.926134, "global_tokens_per_sec": 106440.66451, "grad_norm": 29.5, "loss": 2.090358, "lr_stage": "cosine_decay", "mfu": 61.14362, "stage_name": "permutation_cycle", "status": "training", "step": 1645, "tokens_per_sec_per_gpu": 106440.66451, "vram_gb": 20.709735} {"adam_lr": 9.243610213572285e-06, "dt_ms": 155.32589, "global_tokens_per_sec": 105481.449638, "grad_norm": 37.0, "loss": 2.659655, "lr_stage": "cosine_decay", "mfu": 60.59261, "stage_name": "permutation_cycle", "status": "training", "step": 1646, "tokens_per_sec_per_gpu": 105481.449638, "vram_gb": 20.709735} {"adam_lr": 9.193120495991985e-06, "dt_ms": 153.93734, "global_tokens_per_sec": 106432.916297, "grad_norm": 29.125, "loss": 2.018567, "lr_stage": "cosine_decay", "mfu": 61.139169, "stage_name": "permutation_cycle", "status": "training", "step": 1647, "tokens_per_sec_per_gpu": 106432.916297, "vram_gb": 20.709735} {"adam_lr": 9.142755083243576e-06, "dt_ms": 155.628443, "global_tokens_per_sec": 105276.385916, "grad_norm": 24.25, "loss": 2.439715, "lr_stage": "cosine_decay", "mfu": 60.474814, "stage_name": "permutation_cycle", "status": "training", "step": 1648, "tokens_per_sec_per_gpu": 105276.385916, "vram_gb": 20.709735} {"adam_lr": 9.09251412874882e-06, "dt_ms": 156.195879, "global_tokens_per_sec": 104893.932585, "grad_norm": 39.5, "loss": 2.804034, "lr_stage": "cosine_decay", "mfu": 60.255118, "stage_name": "permutation_cycle", "status": "training", "step": 1649, "tokens_per_sec_per_gpu": 104893.932585, "vram_gb": 20.709735} {"adam_lr": 9.042397785550405e-06, "dt_ms": 155.807734, "global_tokens_per_sec": 105155.242479, "grad_norm": 30.5, "loss": 2.990423, "lr_stage": "cosine_decay", "mfu": 60.405224, "stage_name": "permutation_cycle", "status": "training", "step": 1650, "tokens_per_sec_per_gpu": 105155.242479, "vram_gb": 20.709735} {"adam_lr": 8.992406206311393e-06, "dt_ms": 158.77223, "global_tokens_per_sec": 103191.849007, "grad_norm": 25.75, "loss": 2.082753, "lr_stage": "cosine_decay", "mfu": 59.277375, "stage_name": "permutation_cycle", "status": "training", "step": 1651, "tokens_per_sec_per_gpu": 103191.849007, "vram_gb": 20.709735} {"adam_lr": 8.9425395433148e-06, "dt_ms": 157.494068, "global_tokens_per_sec": 104029.314836, "grad_norm": 24.0, "loss": 2.243668, "lr_stage": "cosine_decay", "mfu": 59.758448, "stage_name": "permutation_cycle", "status": "training", "step": 1652, "tokens_per_sec_per_gpu": 104029.314836, "vram_gb": 20.709735} {"adam_lr": 8.892797948463133e-06, "dt_ms": 157.727242, "global_tokens_per_sec": 103875.524878, "grad_norm": 37.25, "loss": 2.333175, "lr_stage": "cosine_decay", "mfu": 59.670105, "stage_name": "permutation_cycle", "status": "training", "step": 1653, "tokens_per_sec_per_gpu": 103875.524878, "vram_gb": 20.709735} {"adam_lr": 8.843181573277902e-06, "dt_ms": 156.879902, "global_tokens_per_sec": 104436.577299, "grad_norm": 24.125, "loss": 1.323256, "lr_stage": "cosine_decay", "mfu": 59.992395, "stage_name": "permutation_cycle", "status": "training", "step": 1654, "tokens_per_sec_per_gpu": 104436.577299, "vram_gb": 20.709735} {"adam_lr": 8.793690568899216e-06, "dt_ms": 159.94668, "global_tokens_per_sec": 102434.136132, "grad_norm": 33.75, "loss": 2.705134, "lr_stage": "cosine_decay", "mfu": 58.842116, "stage_name": "permutation_cycle", "status": "training", "step": 1655, "tokens_per_sec_per_gpu": 102434.136132, "vram_gb": 20.709735} {"adam_lr": 8.744325086085248e-06, "dt_ms": 157.032728, "global_tokens_per_sec": 104334.938253, "grad_norm": 39.0, "loss": 1.953982, "lr_stage": "cosine_decay", "mfu": 59.93401, "stage_name": "permutation_cycle", "status": "training", "step": 1656, "tokens_per_sec_per_gpu": 104334.938253, "vram_gb": 20.709735} {"adam_lr": 8.695085275211812e-06, "dt_ms": 157.297373, "global_tokens_per_sec": 104159.400163, "grad_norm": 25.75, "loss": 2.328619, "lr_stage": "cosine_decay", "mfu": 59.833174, "stage_name": "permutation_cycle", "status": "training", "step": 1657, "tokens_per_sec_per_gpu": 104159.400163, "vram_gb": 20.709735} {"adam_lr": 8.645971286271904e-06, "dt_ms": 158.190727, "global_tokens_per_sec": 103571.1782, "grad_norm": 35.75, "loss": 2.497397, "lr_stage": "cosine_decay", "mfu": 59.495277, "stage_name": "permutation_cycle", "status": "training", "step": 1658, "tokens_per_sec_per_gpu": 103571.1782, "vram_gb": 20.709735} {"adam_lr": 8.59698326887528e-06, "dt_ms": 157.340288, "global_tokens_per_sec": 104130.99017, "grad_norm": 32.5, "loss": 2.729241, "lr_stage": "cosine_decay", "mfu": 59.816854, "stage_name": "permutation_cycle", "status": "training", "step": 1659, "tokens_per_sec_per_gpu": 104130.99017, "vram_gb": 20.709735} {"adam_lr": 8.548121372247918e-06, "dt_ms": 155.743122, "global_tokens_per_sec": 105198.867077, "grad_norm": 42.75, "loss": 2.763071, "lr_stage": "cosine_decay", "mfu": 60.430284, "stage_name": "permutation_cycle", "status": "training", "step": 1660, "tokens_per_sec_per_gpu": 105198.867077, "vram_gb": 20.709735} {"adam_lr": 8.499385745231631e-06, "dt_ms": 155.642271, "global_tokens_per_sec": 105267.032473, "grad_norm": 24.75, "loss": 1.718997, "lr_stage": "cosine_decay", "mfu": 60.469441, "stage_name": "permutation_cycle", "status": "training", "step": 1661, "tokens_per_sec_per_gpu": 105267.032473, "vram_gb": 20.709735} {"adam_lr": 8.450776536283594e-06, "dt_ms": 157.36413, "global_tokens_per_sec": 104115.213536, "grad_norm": 30.75, "loss": 2.236152, "lr_stage": "cosine_decay", "mfu": 59.807791, "stage_name": "permutation_cycle", "status": "training", "step": 1662, "tokens_per_sec_per_gpu": 104115.213536, "vram_gb": 20.709735} {"adam_lr": 8.40229389347587e-06, "dt_ms": 155.837297, "global_tokens_per_sec": 105135.293471, "grad_norm": 27.5, "loss": 2.634682, "lr_stage": "cosine_decay", "mfu": 60.393765, "stage_name": "permutation_cycle", "status": "training", "step": 1663, "tokens_per_sec_per_gpu": 105135.293471, "vram_gb": 20.709735} {"adam_lr": 8.353937964495029e-06, "dt_ms": 155.468941, "global_tokens_per_sec": 105384.393324, "grad_norm": 27.875, "loss": 1.724124, "lr_stage": "cosine_decay", "mfu": 60.536857, "stage_name": "permutation_cycle", "status": "training", "step": 1664, "tokens_per_sec_per_gpu": 105384.393324, "vram_gb": 20.709735} {"adam_lr": 8.305708896641594e-06, "dt_ms": 155.293703, "global_tokens_per_sec": 105503.311951, "grad_norm": 19.5, "loss": 2.513249, "lr_stage": "cosine_decay", "mfu": 60.605169, "stage_name": "permutation_cycle", "status": "training", "step": 1665, "tokens_per_sec_per_gpu": 105503.311951, "vram_gb": 20.709735} {"adam_lr": 8.257606836829678e-06, "dt_ms": 156.246662, "global_tokens_per_sec": 104859.840048, "grad_norm": 16.125, "loss": 1.895288, "lr_stage": "cosine_decay", "mfu": 60.235534, "stage_name": "permutation_cycle", "status": "training", "step": 1666, "tokens_per_sec_per_gpu": 104859.840048, "vram_gb": 20.709735} {"adam_lr": 8.209631931586498e-06, "dt_ms": 156.298637, "global_tokens_per_sec": 104824.970157, "grad_norm": 25.875, "loss": 2.415206, "lr_stage": "cosine_decay", "mfu": 60.215503, "stage_name": "permutation_cycle", "status": "training", "step": 1667, "tokens_per_sec_per_gpu": 104824.970157, "vram_gb": 20.709735} {"adam_lr": 8.16178432705192e-06, "dt_ms": 156.328201, "global_tokens_per_sec": 104805.146252, "grad_norm": 34.0, "loss": 3.012716, "lr_stage": "cosine_decay", "mfu": 60.204115, "stage_name": "permutation_cycle", "status": "training", "step": 1668, "tokens_per_sec_per_gpu": 104805.146252, "vram_gb": 20.709735} {"adam_lr": 8.114064168978064e-06, "dt_ms": 157.345295, "global_tokens_per_sec": 104127.67668, "grad_norm": 26.75, "loss": 1.925723, "lr_stage": "cosine_decay", "mfu": 59.814951, "stage_name": "permutation_cycle", "status": "training", "step": 1669, "tokens_per_sec_per_gpu": 104127.67668, "vram_gb": 20.709735} {"adam_lr": 8.066471602728803e-06, "dt_ms": 159.170389, "global_tokens_per_sec": 102933.718293, "grad_norm": 26.875, "loss": 1.599957, "lr_stage": "cosine_decay", "mfu": 59.129095, "stage_name": "permutation_cycle", "status": "training", "step": 1670, "tokens_per_sec_per_gpu": 102933.718293, "vram_gb": 20.709735} {"adam_lr": 8.019006773279348e-06, "dt_ms": 158.519745, "global_tokens_per_sec": 103356.209746, "grad_norm": 15.5, "loss": 2.109364, "lr_stage": "cosine_decay", "mfu": 59.37179, "stage_name": "permutation_cycle", "status": "training", "step": 1671, "tokens_per_sec_per_gpu": 103356.209746, "vram_gb": 20.709735} {"adam_lr": 7.971669825215788e-06, "dt_ms": 154.437065, "global_tokens_per_sec": 106088.522123, "grad_norm": 26.0, "loss": 1.678701, "lr_stage": "cosine_decay", "mfu": 60.941336, "stage_name": "permutation_cycle", "status": "training", "step": 1672, "tokens_per_sec_per_gpu": 106088.522123, "vram_gb": 20.709735} {"adam_lr": 7.924460902734699e-06, "dt_ms": 154.37746, "global_tokens_per_sec": 106129.482562, "grad_norm": 39.0, "loss": 1.423462, "lr_stage": "cosine_decay", "mfu": 60.964865, "stage_name": "permutation_cycle", "status": "training", "step": 1673, "tokens_per_sec_per_gpu": 106129.482562, "vram_gb": 20.709735} {"adam_lr": 7.877380149642626e-06, "dt_ms": 155.144215, "global_tokens_per_sec": 105604.969151, "grad_norm": 34.75, "loss": 2.772007, "lr_stage": "cosine_decay", "mfu": 60.663565, "stage_name": "permutation_cycle", "status": "training", "step": 1674, "tokens_per_sec_per_gpu": 105604.969151, "vram_gb": 20.709735} {"adam_lr": 7.830427709355725e-06, "dt_ms": 154.157639, "global_tokens_per_sec": 106280.81848, "grad_norm": 30.125, "loss": 2.166376, "lr_stage": "cosine_decay", "mfu": 61.051798, "stage_name": "permutation_cycle", "status": "training", "step": 1675, "tokens_per_sec_per_gpu": 106280.81848, "vram_gb": 20.709735} {"adam_lr": 7.783603724899257e-06, "dt_ms": 154.917955, "global_tokens_per_sec": 105759.20627, "grad_norm": 38.25, "loss": 2.107283, "lr_stage": "cosine_decay", "mfu": 60.752164, "stage_name": "permutation_cycle", "status": "training", "step": 1676, "tokens_per_sec_per_gpu": 105759.20627, "vram_gb": 20.709735} {"adam_lr": 7.736908338907195e-06, "dt_ms": 154.063463, "global_tokens_per_sec": 106345.785422, "grad_norm": 30.375, "loss": 2.0036, "lr_stage": "cosine_decay", "mfu": 61.089118, "stage_name": "permutation_cycle", "status": "training", "step": 1677, "tokens_per_sec_per_gpu": 106345.785422, "vram_gb": 20.709735} {"adam_lr": 7.690341693621805e-06, "dt_ms": 156.088352, "global_tokens_per_sec": 104966.192344, "grad_norm": 30.625, "loss": 2.263797, "lr_stage": "cosine_decay", "mfu": 60.296627, "stage_name": "permutation_cycle", "status": "training", "step": 1678, "tokens_per_sec_per_gpu": 104966.192344, "vram_gb": 20.709735} {"adam_lr": 7.643903930893154e-06, "dt_ms": 155.814409, "global_tokens_per_sec": 105150.737202, "grad_norm": 29.875, "loss": 2.456645, "lr_stage": "cosine_decay", "mfu": 60.402636, "stage_name": "permutation_cycle", "status": "training", "step": 1679, "tokens_per_sec_per_gpu": 105150.737202, "vram_gb": 20.709735} {"adam_lr": 7.597595192178702e-06, "dt_ms": 156.890154, "global_tokens_per_sec": 104429.752883, "grad_norm": 33.25, "loss": 3.013944, "lr_stage": "cosine_decay", "mfu": 59.988475, "stage_name": "permutation_cycle", "status": "training", "step": 1680, "tokens_per_sec_per_gpu": 104429.752883, "vram_gb": 20.709735} {"adam_lr": 7.5514156185429286e-06, "dt_ms": 153.596163, "global_tokens_per_sec": 106669.331458, "grad_norm": 23.75, "loss": 1.518572, "lr_stage": "cosine_decay", "mfu": 61.274975, "stage_name": "permutation_cycle", "status": "training", "step": 1681, "tokens_per_sec_per_gpu": 106669.331458, "vram_gb": 20.709735} {"adam_lr": 7.505365350656812e-06, "dt_ms": 155.025959, "global_tokens_per_sec": 105685.525857, "grad_norm": 28.625, "loss": 2.091299, "lr_stage": "cosine_decay", "mfu": 60.709839, "stage_name": "permutation_cycle", "status": "training", "step": 1682, "tokens_per_sec_per_gpu": 105685.525857, "vram_gb": 20.709735} {"adam_lr": 7.459444528797438e-06, "dt_ms": 154.160023, "global_tokens_per_sec": 106279.174777, "grad_norm": 22.5, "loss": 1.604556, "lr_stage": "cosine_decay", "mfu": 61.050854, "stage_name": "permutation_cycle", "status": "training", "step": 1683, "tokens_per_sec_per_gpu": 106279.174777, "vram_gb": 20.709735} {"adam_lr": 7.413653292847617e-06, "dt_ms": 154.533625, "global_tokens_per_sec": 106022.233266, "grad_norm": 35.25, "loss": 2.064613, "lr_stage": "cosine_decay", "mfu": 60.903257, "stage_name": "permutation_cycle", "status": "training", "step": 1684, "tokens_per_sec_per_gpu": 106022.233266, "vram_gb": 20.709735} {"adam_lr": 7.367991782295391e-06, "dt_ms": 156.853914, "global_tokens_per_sec": 104453.880397, "grad_norm": 23.25, "loss": 2.116885, "lr_stage": "cosine_decay", "mfu": 60.002335, "stage_name": "permutation_cycle", "status": "training", "step": 1685, "tokens_per_sec_per_gpu": 104453.880397, "vram_gb": 20.709735} {"adam_lr": 7.322460136233622e-06, "dt_ms": 156.737328, "global_tokens_per_sec": 104531.576833, "grad_norm": 15.9375, "loss": 2.329968, "lr_stage": "cosine_decay", "mfu": 60.046967, "stage_name": "permutation_cycle", "status": "training", "step": 1686, "tokens_per_sec_per_gpu": 104531.576833, "vram_gb": 20.709735} {"adam_lr": 7.2770584933596284e-06, "dt_ms": 155.308008, "global_tokens_per_sec": 105493.594249, "grad_norm": 26.125, "loss": 2.419266, "lr_stage": "cosine_decay", "mfu": 60.599587, "stage_name": "permutation_cycle", "status": "training", "step": 1687, "tokens_per_sec_per_gpu": 105493.594249, "vram_gb": 20.709735} {"adam_lr": 7.2317869919746705e-06, "dt_ms": 154.574394, "global_tokens_per_sec": 105994.269504, "grad_norm": 27.625, "loss": 2.100777, "lr_stage": "cosine_decay", "mfu": 60.887194, "stage_name": "permutation_cycle", "status": "training", "step": 1688, "tokens_per_sec_per_gpu": 105994.269504, "vram_gb": 20.709735} {"adam_lr": 7.186645769983591e-06, "dt_ms": 156.566858, "global_tokens_per_sec": 104645.390339, "grad_norm": 31.5, "loss": 2.350097, "lr_stage": "cosine_decay", "mfu": 60.112345, "stage_name": "permutation_cycle", "status": "training", "step": 1689, "tokens_per_sec_per_gpu": 104645.390339, "vram_gb": 20.709735} {"adam_lr": 7.1416349648943894e-06, "dt_ms": 154.491663, "global_tokens_per_sec": 106051.030095, "grad_norm": 20.25, "loss": 2.134754, "lr_stage": "cosine_decay", "mfu": 60.919799, "stage_name": "permutation_cycle", "status": "training", "step": 1690, "tokens_per_sec_per_gpu": 106051.030095, "vram_gb": 20.709735} {"adam_lr": 7.096754713817771e-06, "dt_ms": 153.403044, "global_tokens_per_sec": 106803.61745, "grad_norm": 28.125, "loss": 1.448919, "lr_stage": "cosine_decay", "mfu": 61.352114, "stage_name": "permutation_cycle", "status": "training", "step": 1691, "tokens_per_sec_per_gpu": 106803.61745, "vram_gb": 20.709735} {"adam_lr": 7.052005153466779e-06, "dt_ms": 158.889294, "global_tokens_per_sec": 103115.821221, "grad_norm": 28.125, "loss": 2.217165, "lr_stage": "cosine_decay", "mfu": 59.233702, "stage_name": "permutation_cycle", "status": "training", "step": 1692, "tokens_per_sec_per_gpu": 103115.821221, "vram_gb": 20.709735} {"adam_lr": 7.007386420156331e-06, "dt_ms": 158.82206, "global_tokens_per_sec": 103159.473174, "grad_norm": 22.625, "loss": 1.830512, "lr_stage": "cosine_decay", "mfu": 59.258777, "stage_name": "permutation_cycle", "status": "training", "step": 1693, "tokens_per_sec_per_gpu": 103159.473174, "vram_gb": 20.709735} {"adam_lr": 6.962898649802823e-06, "dt_ms": 159.571409, "global_tokens_per_sec": 102675.034829, "grad_norm": 33.5, "loss": 2.520767, "lr_stage": "cosine_decay", "mfu": 58.980497, "stage_name": "permutation_cycle", "status": "training", "step": 1694, "tokens_per_sec_per_gpu": 102675.034829, "vram_gb": 20.709735} {"adam_lr": 6.918541977923709e-06, "dt_ms": 159.219503, "global_tokens_per_sec": 102901.966467, "grad_norm": 32.75, "loss": 2.566338, "lr_stage": "cosine_decay", "mfu": 59.110856, "stage_name": "permutation_cycle", "status": "training", "step": 1695, "tokens_per_sec_per_gpu": 102901.966467, "vram_gb": 20.709735} {"adam_lr": 6.874316539637127e-06, "dt_ms": 173.288345, "global_tokens_per_sec": 94547.616391, "grad_norm": 29.0, "loss": 1.737195, "lr_stage": "cosine_decay", "mfu": 54.311795, "stage_name": "permutation_cycle", "status": "training", "step": 1696, "tokens_per_sec_per_gpu": 94547.616391, "vram_gb": 20.709735} {"adam_lr": 6.830222469661418e-06, "dt_ms": 154.148102, "global_tokens_per_sec": 106287.393798, "grad_norm": 32.0, "loss": 2.137505, "lr_stage": "cosine_decay", "mfu": 61.055576, "stage_name": "permutation_cycle", "status": "training", "step": 1697, "tokens_per_sec_per_gpu": 106287.393798, "vram_gb": 20.709735} {"adam_lr": 6.786259902314768e-06, "dt_ms": 156.414747, "global_tokens_per_sec": 104747.15645, "grad_norm": 30.375, "loss": 2.183239, "lr_stage": "cosine_decay", "mfu": 60.170804, "stage_name": "permutation_cycle", "status": "training", "step": 1698, "tokens_per_sec_per_gpu": 104747.15645, "vram_gb": 20.709735} {"adam_lr": 6.742428971514786e-06, "dt_ms": 156.187296, "global_tokens_per_sec": 104899.696894, "grad_norm": 25.25, "loss": 2.191558, "lr_stage": "cosine_decay", "mfu": 60.258429, "stage_name": "permutation_cycle", "status": "training", "step": 1699, "tokens_per_sec_per_gpu": 104899.696894, "vram_gb": 20.709735} {"adam_lr": 6.698729810778065e-06, "dt_ms": 155.027151, "global_tokens_per_sec": 105684.71318, "grad_norm": 38.5, "loss": 2.481793, "lr_stage": "cosine_decay", "mfu": 60.709373, "stage_name": "permutation_cycle", "status": "training", "step": 1700, "tokens_per_sec_per_gpu": 105684.71318, "vram_gb": 20.709735} {"adam_lr": 6.655162553219863e-06, "dt_ms": 155.3514, "global_tokens_per_sec": 105464.128166, "grad_norm": 26.875, "loss": 2.206579, "lr_stage": "cosine_decay", "mfu": 60.58266, "stage_name": "permutation_cycle", "status": "training", "step": 1701, "tokens_per_sec_per_gpu": 105464.128166, "vram_gb": 20.709735} {"adam_lr": 6.611727331553586e-06, "dt_ms": 159.887791, "global_tokens_per_sec": 102471.864364, "grad_norm": 34.5, "loss": 2.023909, "lr_stage": "cosine_decay", "mfu": 58.863788, "stage_name": "permutation_cycle", "status": "training", "step": 1702, "tokens_per_sec_per_gpu": 102471.864364, "vram_gb": 20.709735} {"adam_lr": 6.568424278090446e-06, "dt_ms": 159.114838, "global_tokens_per_sec": 102969.655391, "grad_norm": 21.875, "loss": 2.711924, "lr_stage": "cosine_decay", "mfu": 59.149739, "stage_name": "permutation_cycle", "status": "training", "step": 1703, "tokens_per_sec_per_gpu": 102969.655391, "vram_gb": 20.709735} {"adam_lr": 6.52525352473905e-06, "dt_ms": 158.528328, "global_tokens_per_sec": 103350.613816, "grad_norm": 34.25, "loss": 1.89143, "lr_stage": "cosine_decay", "mfu": 59.368576, "stage_name": "permutation_cycle", "status": "training", "step": 1704, "tokens_per_sec_per_gpu": 103350.613816, "vram_gb": 20.709735} {"adam_lr": 6.482215203005015e-06, "dt_ms": 165.290833, "global_tokens_per_sec": 99122.254697, "grad_norm": 18.875, "loss": 2.659485, "lr_stage": "cosine_decay", "mfu": 56.939643, "stage_name": "permutation_cycle", "status": "training", "step": 1705, "tokens_per_sec_per_gpu": 99122.254697, "vram_gb": 20.709735} {"adam_lr": 6.439309443990532e-06, "dt_ms": 155.657053, "global_tokens_per_sec": 105257.035803, "grad_norm": 21.625, "loss": 2.040285, "lr_stage": "cosine_decay", "mfu": 60.463698, "stage_name": "permutation_cycle", "status": "training", "step": 1706, "tokens_per_sec_per_gpu": 105257.035803, "vram_gb": 20.709735} {"adam_lr": 6.396536378393975e-06, "dt_ms": 156.24404, "global_tokens_per_sec": 104861.600153, "grad_norm": 33.0, "loss": 2.54324, "lr_stage": "cosine_decay", "mfu": 60.236545, "stage_name": "permutation_cycle", "status": "training", "step": 1707, "tokens_per_sec_per_gpu": 104861.600153, "vram_gb": 20.709735} {"adam_lr": 6.353896136509524e-06, "dt_ms": 157.13048, "global_tokens_per_sec": 104270.030993, "grad_norm": 29.5, "loss": 2.399874, "lr_stage": "cosine_decay", "mfu": 59.896725, "stage_name": "permutation_cycle", "status": "training", "step": 1708, "tokens_per_sec_per_gpu": 104270.030993, "vram_gb": 20.709735} {"adam_lr": 6.31138884822674e-06, "dt_ms": 154.826164, "global_tokens_per_sec": 105821.907297, "grad_norm": 21.25, "loss": 1.602615, "lr_stage": "cosine_decay", "mfu": 60.788182, "stage_name": "permutation_cycle", "status": "training", "step": 1709, "tokens_per_sec_per_gpu": 105821.907297, "vram_gb": 20.709735} {"adam_lr": 6.269014643030213e-06, "dt_ms": 152.866364, "global_tokens_per_sec": 107178.581489, "grad_norm": 30.625, "loss": 1.158451, "lr_stage": "cosine_decay", "mfu": 61.567508, "stage_name": "permutation_cycle", "status": "training", "step": 1710, "tokens_per_sec_per_gpu": 107178.581489, "vram_gb": 20.709735} {"adam_lr": 6.2267736499991125e-06, "dt_ms": 155.117989, "global_tokens_per_sec": 105622.82395, "grad_norm": 17.125, "loss": 2.100118, "lr_stage": "cosine_decay", "mfu": 60.673821, "stage_name": "permutation_cycle", "status": "training", "step": 1711, "tokens_per_sec_per_gpu": 105622.82395, "vram_gb": 20.709735} {"adam_lr": 6.184665997806832e-06, "dt_ms": 155.555964, "global_tokens_per_sec": 105325.438059, "grad_norm": 31.375, "loss": 1.931607, "lr_stage": "cosine_decay", "mfu": 60.502991, "stage_name": "permutation_cycle", "status": "training", "step": 1712, "tokens_per_sec_per_gpu": 105325.438059, "vram_gb": 20.709735} {"adam_lr": 6.142691814720575e-06, "dt_ms": 156.408548, "global_tokens_per_sec": 104751.307856, "grad_norm": 23.875, "loss": 1.538508, "lr_stage": "cosine_decay", "mfu": 60.173189, "stage_name": "permutation_cycle", "status": "training", "step": 1713, "tokens_per_sec_per_gpu": 104751.307856, "vram_gb": 20.709735} {"adam_lr": 6.100851228600973e-06, "dt_ms": 155.781746, "global_tokens_per_sec": 105172.784553, "grad_norm": 35.0, "loss": 2.409008, "lr_stage": "cosine_decay", "mfu": 60.415301, "stage_name": "permutation_cycle", "status": "training", "step": 1714, "tokens_per_sec_per_gpu": 105172.784553, "vram_gb": 20.709735} {"adam_lr": 6.059144366901736e-06, "dt_ms": 156.549692, "global_tokens_per_sec": 104656.865016, "grad_norm": 34.0, "loss": 2.608378, "lr_stage": "cosine_decay", "mfu": 60.118937, "stage_name": "permutation_cycle", "status": "training", "step": 1715, "tokens_per_sec_per_gpu": 104656.865016, "vram_gb": 20.709735} {"adam_lr": 6.017571356669183e-06, "dt_ms": 155.077219, "global_tokens_per_sec": 105650.592038, "grad_norm": 20.25, "loss": 1.921958, "lr_stage": "cosine_decay", "mfu": 60.689772, "stage_name": "permutation_cycle", "status": "training", "step": 1716, "tokens_per_sec_per_gpu": 105650.592038, "vram_gb": 20.709735} {"adam_lr": 5.9761323245419e-06, "dt_ms": 153.371811, "global_tokens_per_sec": 106825.367077, "grad_norm": 32.75, "loss": 1.482624, "lr_stage": "cosine_decay", "mfu": 61.364608, "stage_name": "permutation_cycle", "status": "training", "step": 1717, "tokens_per_sec_per_gpu": 106825.367077, "vram_gb": 20.709735} {"adam_lr": 5.934827396750392e-06, "dt_ms": 156.665802, "global_tokens_per_sec": 104579.300592, "grad_norm": 24.875, "loss": 2.267047, "lr_stage": "cosine_decay", "mfu": 60.074381, "stage_name": "permutation_cycle", "status": "training", "step": 1718, "tokens_per_sec_per_gpu": 104579.300592, "vram_gb": 20.709735} {"adam_lr": 5.8936566991166175e-06, "dt_ms": 155.318499, "global_tokens_per_sec": 105486.469071, "grad_norm": 36.25, "loss": 1.884533, "lr_stage": "cosine_decay", "mfu": 60.595494, "stage_name": "permutation_cycle", "status": "training", "step": 1719, "tokens_per_sec_per_gpu": 105486.469071, "vram_gb": 20.709735} {"adam_lr": 5.852620357053651e-06, "dt_ms": 154.921293, "global_tokens_per_sec": 105756.927633, "grad_norm": 27.25, "loss": 2.256744, "lr_stage": "cosine_decay", "mfu": 60.750855, "stage_name": "permutation_cycle", "status": "training", "step": 1720, "tokens_per_sec_per_gpu": 105756.927633, "vram_gb": 20.709735} {"adam_lr": 5.811718495565327e-06, "dt_ms": 156.107426, "global_tokens_per_sec": 104953.367385, "grad_norm": 27.125, "loss": 2.736928, "lr_stage": "cosine_decay", "mfu": 60.289259, "stage_name": "permutation_cycle", "status": "training", "step": 1721, "tokens_per_sec_per_gpu": 104953.367385, "vram_gb": 20.709735} {"adam_lr": 5.770951239245803e-06, "dt_ms": 153.141499, "global_tokens_per_sec": 106986.023733, "grad_norm": 18.625, "loss": 1.687761, "lr_stage": "cosine_decay", "mfu": 61.456896, "stage_name": "permutation_cycle", "status": "training", "step": 1722, "tokens_per_sec_per_gpu": 106986.023733, "vram_gb": 20.709735} {"adam_lr": 5.730318712279203e-06, "dt_ms": 154.992104, "global_tokens_per_sec": 105708.611096, "grad_norm": 25.375, "loss": 2.421788, "lr_stage": "cosine_decay", "mfu": 60.7231, "stage_name": "permutation_cycle", "status": "training", "step": 1723, "tokens_per_sec_per_gpu": 105708.611096, "vram_gb": 20.709735} {"adam_lr": 5.689821038439263e-06, "dt_ms": 155.441999, "global_tokens_per_sec": 105402.658609, "grad_norm": 34.5, "loss": 2.119732, "lr_stage": "cosine_decay", "mfu": 60.54735, "stage_name": "permutation_cycle", "status": "training", "step": 1724, "tokens_per_sec_per_gpu": 105402.658609, "vram_gb": 20.709735} {"adam_lr": 5.649458341088915e-06, "dt_ms": 155.651093, "global_tokens_per_sec": 105261.06649, "grad_norm": 41.5, "loss": 1.969839, "lr_stage": "cosine_decay", "mfu": 60.466014, "stage_name": "permutation_cycle", "status": "training", "step": 1725, "tokens_per_sec_per_gpu": 105261.06649, "vram_gb": 20.709735} {"adam_lr": 5.6092307431799384e-06, "dt_ms": 155.886412, "global_tokens_per_sec": 105102.169104, "grad_norm": 32.25, "loss": 2.500442, "lr_stage": "cosine_decay", "mfu": 60.374737, "stage_name": "permutation_cycle", "status": "training", "step": 1726, "tokens_per_sec_per_gpu": 105102.169104, "vram_gb": 20.709735} {"adam_lr": 5.569138367252552e-06, "dt_ms": 155.844927, "global_tokens_per_sec": 105130.146568, "grad_norm": 26.5, "loss": 2.128451, "lr_stage": "cosine_decay", "mfu": 60.390808, "stage_name": "permutation_cycle", "status": "training", "step": 1727, "tokens_per_sec_per_gpu": 105130.146568, "vram_gb": 20.709735} {"adam_lr": 5.529181335435124e-06, "dt_ms": 156.585455, "global_tokens_per_sec": 104632.962277, "grad_norm": 27.625, "loss": 2.272609, "lr_stage": "cosine_decay", "mfu": 60.105206, "stage_name": "permutation_cycle", "status": "training", "step": 1728, "tokens_per_sec_per_gpu": 104632.962277, "vram_gb": 20.709735} {"adam_lr": 5.4893597694436746e-06, "dt_ms": 155.892372, "global_tokens_per_sec": 105098.150577, "grad_norm": 32.75, "loss": 1.691399, "lr_stage": "cosine_decay", "mfu": 60.372428, "stage_name": "permutation_cycle", "status": "training", "step": 1729, "tokens_per_sec_per_gpu": 105098.150577, "vram_gb": 20.709735} {"adam_lr": 5.449673790581611e-06, "dt_ms": 155.323982, "global_tokens_per_sec": 105482.74493, "grad_norm": 28.625, "loss": 3.139616, "lr_stage": "cosine_decay", "mfu": 60.593354, "stage_name": "permutation_cycle", "status": "training", "step": 1730, "tokens_per_sec_per_gpu": 105482.74493, "vram_gb": 20.709735} {"adam_lr": 5.410123519739302e-06, "dt_ms": 155.385017, "global_tokens_per_sec": 105441.311361, "grad_norm": 26.25, "loss": 1.764843, "lr_stage": "cosine_decay", "mfu": 60.569553, "stage_name": "permutation_cycle", "status": "training", "step": 1731, "tokens_per_sec_per_gpu": 105441.311361, "vram_gb": 20.709735} {"adam_lr": 5.370709077393721e-06, "dt_ms": 154.362917, "global_tokens_per_sec": 106139.48171, "grad_norm": 36.0, "loss": 2.704725, "lr_stage": "cosine_decay", "mfu": 60.970609, "stage_name": "permutation_cycle", "status": "training", "step": 1732, "tokens_per_sec_per_gpu": 106139.48171, "vram_gb": 20.709735} {"adam_lr": 5.331430583608121e-06, "dt_ms": 154.694319, "global_tokens_per_sec": 105912.098971, "grad_norm": 23.75, "loss": 2.635557, "lr_stage": "cosine_decay", "mfu": 60.839992, "stage_name": "permutation_cycle", "status": "training", "step": 1733, "tokens_per_sec_per_gpu": 105912.098971, "vram_gb": 20.709735} {"adam_lr": 5.292288158031594e-06, "dt_ms": 155.248165, "global_tokens_per_sec": 105534.258561, "grad_norm": 36.5, "loss": 1.946501, "lr_stage": "cosine_decay", "mfu": 60.622946, "stage_name": "permutation_cycle", "status": "training", "step": 1734, "tokens_per_sec_per_gpu": 105534.258561, "vram_gb": 20.709735} {"adam_lr": 5.2532819198987506e-06, "dt_ms": 156.695604, "global_tokens_per_sec": 104559.410397, "grad_norm": 25.5, "loss": 2.240941, "lr_stage": "cosine_decay", "mfu": 60.062955, "stage_name": "permutation_cycle", "status": "training", "step": 1735, "tokens_per_sec_per_gpu": 104559.410397, "vram_gb": 20.709735} {"adam_lr": 5.214411988029355e-06, "dt_ms": 154.734612, "global_tokens_per_sec": 105884.519565, "grad_norm": 30.625, "loss": 2.76507, "lr_stage": "cosine_decay", "mfu": 60.824149, "stage_name": "permutation_cycle", "status": "training", "step": 1736, "tokens_per_sec_per_gpu": 105884.519565, "vram_gb": 20.709735} {"adam_lr": 5.175678480827972e-06, "dt_ms": 154.852867, "global_tokens_per_sec": 105803.659332, "grad_norm": 24.375, "loss": 2.330415, "lr_stage": "cosine_decay", "mfu": 60.7777, "stage_name": "permutation_cycle", "status": "training", "step": 1737, "tokens_per_sec_per_gpu": 105803.659332, "vram_gb": 20.709735} {"adam_lr": 5.137081516283581e-06, "dt_ms": 156.402588, "global_tokens_per_sec": 104755.299902, "grad_norm": 33.5, "loss": 2.728125, "lr_stage": "cosine_decay", "mfu": 60.175482, "stage_name": "permutation_cycle", "status": "training", "step": 1738, "tokens_per_sec_per_gpu": 104755.299902, "vram_gb": 20.709735} {"adam_lr": 5.098621211969223e-06, "dt_ms": 156.273842, "global_tokens_per_sec": 104841.602441, "grad_norm": 24.125, "loss": 2.561325, "lr_stage": "cosine_decay", "mfu": 60.225057, "stage_name": "permutation_cycle", "status": "training", "step": 1739, "tokens_per_sec_per_gpu": 104841.602441, "vram_gb": 20.709735} {"adam_lr": 5.060297685041659e-06, "dt_ms": 155.275345, "global_tokens_per_sec": 105515.785626, "grad_norm": 22.5, "loss": 1.704532, "lr_stage": "cosine_decay", "mfu": 60.612334, "stage_name": "permutation_cycle", "status": "training", "step": 1740, "tokens_per_sec_per_gpu": 105515.785626, "vram_gb": 20.709735} {"adam_lr": 5.0221110522409856e-06, "dt_ms": 154.900312, "global_tokens_per_sec": 105771.252128, "grad_norm": 30.25, "loss": 2.325922, "lr_stage": "cosine_decay", "mfu": 60.759084, "stage_name": "permutation_cycle", "status": "training", "step": 1741, "tokens_per_sec_per_gpu": 105771.252128, "vram_gb": 20.709735} {"adam_lr": 4.984061429890324e-06, "dt_ms": 155.249119, "global_tokens_per_sec": 105533.610278, "grad_norm": 37.5, "loss": 2.5478, "lr_stage": "cosine_decay", "mfu": 60.622573, "stage_name": "permutation_cycle", "status": "training", "step": 1742, "tokens_per_sec_per_gpu": 105533.610278, "vram_gb": 20.709735} {"adam_lr": 4.9461489338954226e-06, "dt_ms": 154.943466, "global_tokens_per_sec": 105741.793463, "grad_norm": 26.375, "loss": 1.641205, "lr_stage": "cosine_decay", "mfu": 60.742162, "stage_name": "permutation_cycle", "status": "training", "step": 1743, "tokens_per_sec_per_gpu": 105741.793463, "vram_gb": 20.709735} {"adam_lr": 4.908373679744316e-06, "dt_ms": 154.469252, "global_tokens_per_sec": 106066.416629, "grad_norm": 12.125, "loss": 1.681404, "lr_stage": "cosine_decay", "mfu": 60.928638, "stage_name": "permutation_cycle", "status": "training", "step": 1744, "tokens_per_sec_per_gpu": 106066.416629, "vram_gb": 20.709735} {"adam_lr": 4.87073578250698e-06, "dt_ms": 153.996706, "global_tokens_per_sec": 106391.886064, "grad_norm": 25.5, "loss": 1.90717, "lr_stage": "cosine_decay", "mfu": 61.1156, "stage_name": "permutation_cycle", "status": "training", "step": 1745, "tokens_per_sec_per_gpu": 106391.886064, "vram_gb": 20.709735} {"adam_lr": 4.833235356834959e-06, "dt_ms": 154.925108, "global_tokens_per_sec": 105754.323597, "grad_norm": 30.625, "loss": 2.112576, "lr_stage": "cosine_decay", "mfu": 60.749359, "stage_name": "permutation_cycle", "status": "training", "step": 1746, "tokens_per_sec_per_gpu": 105754.323597, "vram_gb": 20.709735} {"adam_lr": 4.795872516961087e-06, "dt_ms": 155.602217, "global_tokens_per_sec": 105294.129771, "grad_norm": 37.5, "loss": 2.219508, "lr_stage": "cosine_decay", "mfu": 60.485006, "stage_name": "permutation_cycle", "status": "training", "step": 1747, "tokens_per_sec_per_gpu": 105294.129771, "vram_gb": 20.709735} {"adam_lr": 4.758647376699032e-06, "dt_ms": 156.427145, "global_tokens_per_sec": 104738.854625, "grad_norm": 29.25, "loss": 2.518731, "lr_stage": "cosine_decay", "mfu": 60.166035, "stage_name": "permutation_cycle", "status": "training", "step": 1748, "tokens_per_sec_per_gpu": 104738.854625, "vram_gb": 20.709735} {"adam_lr": 4.7215600494430305e-06, "dt_ms": 154.10924, "global_tokens_per_sec": 106314.196636, "grad_norm": 35.0, "loss": 2.109256, "lr_stage": "cosine_decay", "mfu": 61.070972, "stage_name": "permutation_cycle", "status": "training", "step": 1749, "tokens_per_sec_per_gpu": 106314.196636, "vram_gb": 20.709735} {"adam_lr": 4.684610648167503e-06, "dt_ms": 155.230999, "global_tokens_per_sec": 105545.929011, "grad_norm": 37.0, "loss": 2.387697, "lr_stage": "cosine_decay", "mfu": 60.62965, "stage_name": "permutation_cycle", "status": "training", "step": 1750, "tokens_per_sec_per_gpu": 105545.929011, "vram_gb": 20.709735} {"adam_lr": 4.647799285426757e-06, "dt_ms": 154.843807, "global_tokens_per_sec": 105809.8499, "grad_norm": 23.5, "loss": 2.171345, "lr_stage": "cosine_decay", "mfu": 60.781256, "stage_name": "permutation_cycle", "status": "training", "step": 1751, "tokens_per_sec_per_gpu": 105809.8499, "vram_gb": 20.709735} {"adam_lr": 4.611126073354571e-06, "dt_ms": 158.136129, "global_tokens_per_sec": 103606.937038, "grad_norm": 26.875, "loss": 1.838271, "lr_stage": "cosine_decay", "mfu": 59.515818, "stage_name": "permutation_cycle", "status": "training", "step": 1752, "tokens_per_sec_per_gpu": 103606.937038, "vram_gb": 20.709735} {"adam_lr": 4.574591123663918e-06, "dt_ms": 154.162884, "global_tokens_per_sec": 106277.202401, "grad_norm": 24.875, "loss": 1.904356, "lr_stage": "cosine_decay", "mfu": 61.049721, "stage_name": "permutation_cycle", "status": "training", "step": 1753, "tokens_per_sec_per_gpu": 106277.202401, "vram_gb": 20.709735} {"adam_lr": 4.538194547646574e-06, "dt_ms": 154.206753, "global_tokens_per_sec": 106246.968469, "grad_norm": 31.125, "loss": 1.724336, "lr_stage": "cosine_decay", "mfu": 61.032354, "stage_name": "permutation_cycle", "status": "training", "step": 1754, "tokens_per_sec_per_gpu": 106246.968469, "vram_gb": 20.709735} {"adam_lr": 4.501936456172845e-06, "dt_ms": 155.575752, "global_tokens_per_sec": 105312.040997, "grad_norm": 29.75, "loss": 2.286958, "lr_stage": "cosine_decay", "mfu": 60.495295, "stage_name": "permutation_cycle", "status": "training", "step": 1755, "tokens_per_sec_per_gpu": 105312.040997, "vram_gb": 20.709735} {"adam_lr": 4.465816959691149e-06, "dt_ms": 155.246496, "global_tokens_per_sec": 105535.393075, "grad_norm": 33.75, "loss": 2.649962, "lr_stage": "cosine_decay", "mfu": 60.623597, "stage_name": "permutation_cycle", "status": "training", "step": 1756, "tokens_per_sec_per_gpu": 105535.393075, "vram_gb": 20.709735} {"adam_lr": 4.429836168227735e-06, "dt_ms": 154.583454, "global_tokens_per_sec": 105988.057338, "grad_norm": 26.5, "loss": 1.761549, "lr_stage": "cosine_decay", "mfu": 60.883625, "stage_name": "permutation_cycle", "status": "training", "step": 1757, "tokens_per_sec_per_gpu": 105988.057338, "vram_gb": 20.709735} {"adam_lr": 4.3939941913863525e-06, "dt_ms": 153.833151, "global_tokens_per_sec": 106505.001737, "grad_norm": 21.125, "loss": 1.784785, "lr_stage": "cosine_decay", "mfu": 61.180578, "stage_name": "permutation_cycle", "status": "training", "step": 1758, "tokens_per_sec_per_gpu": 106505.001737, "vram_gb": 20.709735} {"adam_lr": 4.358291138347864e-06, "dt_ms": 155.180931, "global_tokens_per_sec": 105579.982571, "grad_norm": 21.625, "loss": 2.17178, "lr_stage": "cosine_decay", "mfu": 60.649211, "stage_name": "permutation_cycle", "status": "training", "step": 1759, "tokens_per_sec_per_gpu": 105579.982571, "vram_gb": 20.709735} {"adam_lr": 4.322727117869951e-06, "dt_ms": 156.607866, "global_tokens_per_sec": 104617.988792, "grad_norm": 31.5, "loss": 2.138337, "lr_stage": "cosine_decay", "mfu": 60.096605, "stage_name": "permutation_cycle", "status": "training", "step": 1760, "tokens_per_sec_per_gpu": 104617.988792, "vram_gb": 20.709735} {"adam_lr": 4.287302238286811e-06, "dt_ms": 155.982971, "global_tokens_per_sec": 105037.106774, "grad_norm": 20.625, "loss": 2.146058, "lr_stage": "cosine_decay", "mfu": 60.337363, "stage_name": "permutation_cycle", "status": "training", "step": 1761, "tokens_per_sec_per_gpu": 105037.106774, "vram_gb": 20.709735} {"adam_lr": 4.2520166075087635e-06, "dt_ms": 154.562712, "global_tokens_per_sec": 106002.281004, "grad_norm": 24.75, "loss": 2.032844, "lr_stage": "cosine_decay", "mfu": 60.891796, "stage_name": "permutation_cycle", "status": "training", "step": 1762, "tokens_per_sec_per_gpu": 106002.281004, "vram_gb": 20.709735} {"adam_lr": 4.216870333021949e-06, "dt_ms": 155.716896, "global_tokens_per_sec": 105216.584808, "grad_norm": 29.625, "loss": 2.520243, "lr_stage": "cosine_decay", "mfu": 60.440462, "stage_name": "permutation_cycle", "status": "training", "step": 1763, "tokens_per_sec_per_gpu": 105216.584808, "vram_gb": 20.709735} {"adam_lr": 4.181863521888019e-06, "dt_ms": 155.319691, "global_tokens_per_sec": 105485.659453, "grad_norm": 30.75, "loss": 2.567419, "lr_stage": "cosine_decay", "mfu": 60.595028, "stage_name": "permutation_cycle", "status": "training", "step": 1764, "tokens_per_sec_per_gpu": 105485.659453, "vram_gb": 20.709735} {"adam_lr": 4.146996280743798e-06, "dt_ms": 156.294346, "global_tokens_per_sec": 104827.848444, "grad_norm": 31.0, "loss": 2.165675, "lr_stage": "cosine_decay", "mfu": 60.217156, "stage_name": "permutation_cycle", "status": "training", "step": 1765, "tokens_per_sec_per_gpu": 104827.848444, "vram_gb": 20.709735} {"adam_lr": 4.112268715800943e-06, "dt_ms": 156.226397, "global_tokens_per_sec": 104873.442393, "grad_norm": 35.25, "loss": 2.772412, "lr_stage": "cosine_decay", "mfu": 60.243347, "stage_name": "permutation_cycle", "status": "training", "step": 1766, "tokens_per_sec_per_gpu": 104873.442393, "vram_gb": 20.709735} {"adam_lr": 4.077680932845645e-06, "dt_ms": 155.867815, "global_tokens_per_sec": 105114.708884, "grad_norm": 16.75, "loss": 2.474469, "lr_stage": "cosine_decay", "mfu": 60.38194, "stage_name": "permutation_cycle", "status": "training", "step": 1767, "tokens_per_sec_per_gpu": 105114.708884, "vram_gb": 20.709735} {"adam_lr": 4.043233037238281e-06, "dt_ms": 154.266357, "global_tokens_per_sec": 106205.91731, "grad_norm": 34.75, "loss": 2.323406, "lr_stage": "cosine_decay", "mfu": 61.008772, "stage_name": "permutation_cycle", "status": "training", "step": 1768, "tokens_per_sec_per_gpu": 106205.91731, "vram_gb": 20.709735} {"adam_lr": 4.008925133913116e-06, "dt_ms": 155.042887, "global_tokens_per_sec": 105673.987018, "grad_norm": 32.25, "loss": 2.463874, "lr_stage": "cosine_decay", "mfu": 60.703211, "stage_name": "permutation_cycle", "status": "training", "step": 1769, "tokens_per_sec_per_gpu": 105673.987018, "vram_gb": 20.709735} {"adam_lr": 3.974757327377981e-06, "dt_ms": 154.411077, "global_tokens_per_sec": 106106.376986, "grad_norm": 26.5, "loss": 2.330747, "lr_stage": "cosine_decay", "mfu": 60.951593, "stage_name": "permutation_cycle", "status": "training", "step": 1770, "tokens_per_sec_per_gpu": 106106.376986, "vram_gb": 20.709735} {"adam_lr": 3.940729721713943e-06, "dt_ms": 153.70512, "global_tokens_per_sec": 106593.716532, "grad_norm": 27.375, "loss": 2.197314, "lr_stage": "cosine_decay", "mfu": 61.231539, "stage_name": "permutation_cycle", "status": "training", "step": 1771, "tokens_per_sec_per_gpu": 106593.716532, "vram_gb": 20.709735} {"adam_lr": 3.90684242057498e-06, "dt_ms": 156.058073, "global_tokens_per_sec": 104986.558404, "grad_norm": 26.75, "loss": 2.824237, "lr_stage": "cosine_decay", "mfu": 60.308326, "stage_name": "permutation_cycle", "status": "training", "step": 1772, "tokens_per_sec_per_gpu": 104986.558404, "vram_gb": 20.709735} {"adam_lr": 3.873095527187681e-06, "dt_ms": 155.180216, "global_tokens_per_sec": 105580.46921, "grad_norm": 30.125, "loss": 2.262233, "lr_stage": "cosine_decay", "mfu": 60.649491, "stage_name": "permutation_cycle", "status": "training", "step": 1773, "tokens_per_sec_per_gpu": 105580.46921, "vram_gb": 20.709735} {"adam_lr": 3.839489144350955e-06, "dt_ms": 156.383038, "global_tokens_per_sec": 104768.395952, "grad_norm": 19.875, "loss": 1.722452, "lr_stage": "cosine_decay", "mfu": 60.183005, "stage_name": "permutation_cycle", "status": "training", "step": 1774, "tokens_per_sec_per_gpu": 104768.395952, "vram_gb": 20.709735} {"adam_lr": 3.8060233744356633e-06, "dt_ms": 155.801773, "global_tokens_per_sec": 105159.265373, "grad_norm": 22.0, "loss": 2.396294, "lr_stage": "cosine_decay", "mfu": 60.407535, "stage_name": "permutation_cycle", "status": "training", "step": 1775, "tokens_per_sec_per_gpu": 105159.265373, "vram_gb": 20.709735} {"adam_lr": 3.772698319384349e-06, "dt_ms": 154.466152, "global_tokens_per_sec": 106068.544905, "grad_norm": 23.25, "loss": 2.302058, "lr_stage": "cosine_decay", "mfu": 60.92986, "stage_name": "permutation_cycle", "status": "training", "step": 1776, "tokens_per_sec_per_gpu": 106068.544905, "vram_gb": 20.709735} {"adam_lr": 3.7395140807108984e-06, "dt_ms": 158.095837, "global_tokens_per_sec": 103633.342587, "grad_norm": 25.375, "loss": 2.907844, "lr_stage": "cosine_decay", "mfu": 59.530986, "stage_name": "permutation_cycle", "status": "training", "step": 1777, "tokens_per_sec_per_gpu": 103633.342587, "vram_gb": 20.709735} {"adam_lr": 3.7064707595002635e-06, "dt_ms": 155.585766, "global_tokens_per_sec": 105305.26306, "grad_norm": 29.125, "loss": 1.826253, "lr_stage": "cosine_decay", "mfu": 60.491402, "stage_name": "permutation_cycle", "status": "training", "step": 1778, "tokens_per_sec_per_gpu": 105305.26306, "vram_gb": 20.709735} {"adam_lr": 3.6735684564081384e-06, "dt_ms": 158.633709, "global_tokens_per_sec": 103281.957587, "grad_norm": 34.5, "loss": 1.792919, "lr_stage": "cosine_decay", "mfu": 59.329137, "stage_name": "permutation_cycle", "status": "training", "step": 1779, "tokens_per_sec_per_gpu": 103281.957587, "vram_gb": 20.709735} {"adam_lr": 3.6408072716606346e-06, "dt_ms": 157.933712, "global_tokens_per_sec": 103739.725939, "grad_norm": 33.0, "loss": 3.014399, "lr_stage": "cosine_decay", "mfu": 59.592097, "stage_name": "permutation_cycle", "status": "training", "step": 1780, "tokens_per_sec_per_gpu": 103739.725939, "vram_gb": 20.709735} {"adam_lr": 3.608187305054006e-06, "dt_ms": 156.241179, "global_tokens_per_sec": 104863.520334, "grad_norm": 34.0, "loss": 2.649138, "lr_stage": "cosine_decay", "mfu": 60.237648, "stage_name": "permutation_cycle", "status": "training", "step": 1781, "tokens_per_sec_per_gpu": 104863.520334, "vram_gb": 20.709735} {"adam_lr": 3.575708655954324e-06, "dt_ms": 155.902147, "global_tokens_per_sec": 105091.560857, "grad_norm": 23.375, "loss": 2.597362, "lr_stage": "cosine_decay", "mfu": 60.368643, "stage_name": "permutation_cycle", "status": "training", "step": 1782, "tokens_per_sec_per_gpu": 105091.560857, "vram_gb": 20.709735} {"adam_lr": 3.5433714232971925e-06, "dt_ms": 155.506611, "global_tokens_per_sec": 105358.864863, "grad_norm": 27.625, "loss": 2.03588, "lr_stage": "cosine_decay", "mfu": 60.522193, "stage_name": "permutation_cycle", "status": "training", "step": 1783, "tokens_per_sec_per_gpu": 105358.864863, "vram_gb": 20.709735} {"adam_lr": 3.511175705587433e-06, "dt_ms": 156.738997, "global_tokens_per_sec": 104530.463798, "grad_norm": 23.875, "loss": 2.667096, "lr_stage": "cosine_decay", "mfu": 60.046327, "stage_name": "permutation_cycle", "status": "training", "step": 1784, "tokens_per_sec_per_gpu": 104530.463798, "vram_gb": 20.709735} {"adam_lr": 3.479121600898777e-06, "dt_ms": 156.719446, "global_tokens_per_sec": 104543.503688, "grad_norm": 48.25, "loss": 2.730047, "lr_stage": "cosine_decay", "mfu": 60.053818, "stage_name": "permutation_cycle", "status": "training", "step": 1785, "tokens_per_sec_per_gpu": 104543.503688, "vram_gb": 20.709735} {"adam_lr": 3.4472092068735916e-06, "dt_ms": 154.211521, "global_tokens_per_sec": 106243.683208, "grad_norm": 42.5, "loss": 2.217891, "lr_stage": "cosine_decay", "mfu": 61.030467, "stage_name": "permutation_cycle", "status": "training", "step": 1786, "tokens_per_sec_per_gpu": 106243.683208, "vram_gb": 20.709735} {"adam_lr": 3.4154386207225554e-06, "dt_ms": 156.403542, "global_tokens_per_sec": 104754.661155, "grad_norm": 41.25, "loss": 2.116613, "lr_stage": "cosine_decay", "mfu": 60.175115, "stage_name": "permutation_cycle", "status": "training", "step": 1787, "tokens_per_sec_per_gpu": 104754.661155, "vram_gb": 20.709735} {"adam_lr": 3.3838099392243916e-06, "dt_ms": 154.630661, "global_tokens_per_sec": 105955.70046, "grad_norm": 27.0, "loss": 2.191053, "lr_stage": "cosine_decay", "mfu": 60.865038, "stage_name": "permutation_cycle", "status": "training", "step": 1788, "tokens_per_sec_per_gpu": 105955.70046, "vram_gb": 20.709735} {"adam_lr": 3.3523232587255536e-06, "dt_ms": 156.007528, "global_tokens_per_sec": 105020.572904, "grad_norm": 30.375, "loss": 2.808538, "lr_stage": "cosine_decay", "mfu": 60.327865, "stage_name": "permutation_cycle", "status": "training", "step": 1789, "tokens_per_sec_per_gpu": 105020.572904, "vram_gb": 20.709735} {"adam_lr": 3.3209786751399187e-06, "dt_ms": 155.964613, "global_tokens_per_sec": 105049.470447, "grad_norm": 25.75, "loss": 1.727232, "lr_stage": "cosine_decay", "mfu": 60.344465, "stage_name": "permutation_cycle", "status": "training", "step": 1790, "tokens_per_sec_per_gpu": 105049.470447, "vram_gb": 20.709735} {"adam_lr": 3.2897762839485256e-06, "dt_ms": 155.538559, "global_tokens_per_sec": 105337.223834, "grad_norm": 33.25, "loss": 2.817827, "lr_stage": "cosine_decay", "mfu": 60.509761, "stage_name": "permutation_cycle", "status": "training", "step": 1791, "tokens_per_sec_per_gpu": 105337.223834, "vram_gb": 20.709735} {"adam_lr": 3.258716180199278e-06, "dt_ms": 154.698133, "global_tokens_per_sec": 105909.487288, "grad_norm": 27.0, "loss": 2.889905, "lr_stage": "cosine_decay", "mfu": 60.838491, "stage_name": "permutation_cycle", "status": "training", "step": 1792, "tokens_per_sec_per_gpu": 105909.487288, "vram_gb": 20.709735} {"adam_lr": 3.2277984585066366e-06, "dt_ms": 157.995701, "global_tokens_per_sec": 103699.024171, "grad_norm": 28.125, "loss": 2.262138, "lr_stage": "cosine_decay", "mfu": 59.568716, "stage_name": "permutation_cycle", "status": "training", "step": 1793, "tokens_per_sec_per_gpu": 103699.024171, "vram_gb": 20.709735} {"adam_lr": 3.197023213051337e-06, "dt_ms": 155.185938, "global_tokens_per_sec": 105576.576226, "grad_norm": 24.75, "loss": 2.482671, "lr_stage": "cosine_decay", "mfu": 60.647255, "stage_name": "permutation_cycle", "status": "training", "step": 1794, "tokens_per_sec_per_gpu": 105576.576226, "vram_gb": 20.709735} {"adam_lr": 3.1663905375801216e-06, "dt_ms": 155.512094, "global_tokens_per_sec": 105355.149726, "grad_norm": 26.125, "loss": 2.203957, "lr_stage": "cosine_decay", "mfu": 60.520059, "stage_name": "permutation_cycle", "status": "training", "step": 1795, "tokens_per_sec_per_gpu": 105355.149726, "vram_gb": 20.709735} {"adam_lr": 3.1359005254054273e-06, "dt_ms": 154.780149, "global_tokens_per_sec": 105853.367226, "grad_norm": 31.875, "loss": 2.603618, "lr_stage": "cosine_decay", "mfu": 60.806254, "stage_name": "permutation_cycle", "status": "training", "step": 1796, "tokens_per_sec_per_gpu": 105853.367226, "vram_gb": 20.709735} {"adam_lr": 3.105553269405115e-06, "dt_ms": 155.657768, "global_tokens_per_sec": 105256.552142, "grad_norm": 20.0, "loss": 2.202047, "lr_stage": "cosine_decay", "mfu": 60.46342, "stage_name": "permutation_cycle", "status": "training", "step": 1797, "tokens_per_sec_per_gpu": 105256.552142, "vram_gb": 20.709735} {"adam_lr": 3.0753488620222037e-06, "dt_ms": 156.48222, "global_tokens_per_sec": 104701.991267, "grad_norm": 14.8125, "loss": 1.908581, "lr_stage": "cosine_decay", "mfu": 60.144859, "stage_name": "permutation_cycle", "status": "training", "step": 1798, "tokens_per_sec_per_gpu": 104701.991267, "vram_gb": 20.709735} {"adam_lr": 3.0452873952645453e-06, "dt_ms": 154.08349, "global_tokens_per_sec": 106331.963019, "grad_norm": 36.0, "loss": 2.233182, "lr_stage": "cosine_decay", "mfu": 61.081178, "stage_name": "permutation_cycle", "status": "training", "step": 1799, "tokens_per_sec_per_gpu": 106331.963019, "vram_gb": 20.709735} {"adam_lr": 3.0153689607045845e-06, "dt_ms": 156.164885, "global_tokens_per_sec": 104914.751132, "grad_norm": 32.75, "loss": 2.404976, "lr_stage": "cosine_decay", "mfu": 60.267077, "stage_name": "permutation_cycle", "status": "training", "step": 1800, "tokens_per_sec_per_gpu": 104914.751132, "vram_gb": 20.709735} {"adam_lr": 2.9855936494790513e-06, "dt_ms": 155.214071, "global_tokens_per_sec": 105557.439899, "grad_norm": 28.5, "loss": 2.894571, "lr_stage": "cosine_decay", "mfu": 60.636262, "stage_name": "permutation_cycle", "status": "training", "step": 1801, "tokens_per_sec_per_gpu": 105557.439899, "vram_gb": 20.709735} {"adam_lr": 2.9559615522887273e-06, "dt_ms": 155.699968, "global_tokens_per_sec": 105228.023967, "grad_norm": 37.0, "loss": 3.206929, "lr_stage": "cosine_decay", "mfu": 60.447033, "stage_name": "permutation_cycle", "status": "training", "step": 1802, "tokens_per_sec_per_gpu": 105228.023967, "vram_gb": 20.709735} {"adam_lr": 2.9264727593981023e-06, "dt_ms": 154.656172, "global_tokens_per_sec": 105938.222894, "grad_norm": 27.875, "loss": 1.977587, "lr_stage": "cosine_decay", "mfu": 60.854998, "stage_name": "permutation_cycle", "status": "training", "step": 1803, "tokens_per_sec_per_gpu": 105938.222894, "vram_gb": 20.709735} {"adam_lr": 2.8971273606351658e-06, "dt_ms": 156.069517, "global_tokens_per_sec": 104978.860066, "grad_norm": 30.625, "loss": 2.067256, "lr_stage": "cosine_decay", "mfu": 60.303903, "stage_name": "permutation_cycle", "status": "training", "step": 1804, "tokens_per_sec_per_gpu": 104978.860066, "vram_gb": 20.709735} {"adam_lr": 2.8679254453910785e-06, "dt_ms": 152.677298, "global_tokens_per_sec": 107311.304682, "grad_norm": 35.5, "loss": 1.808509, "lr_stage": "cosine_decay", "mfu": 61.643749, "stage_name": "permutation_cycle", "status": "training", "step": 1805, "tokens_per_sec_per_gpu": 107311.304682, "vram_gb": 20.709735} {"adam_lr": 2.8388671026199522e-06, "dt_ms": 156.899214, "global_tokens_per_sec": 104423.722746, "grad_norm": 24.125, "loss": 2.201576, "lr_stage": "cosine_decay", "mfu": 59.985011, "stage_name": "permutation_cycle", "status": "training", "step": 1806, "tokens_per_sec_per_gpu": 104423.722746, "vram_gb": 20.709735} {"adam_lr": 2.8099524208385296e-06, "dt_ms": 154.768229, "global_tokens_per_sec": 105861.520517, "grad_norm": 23.875, "loss": 2.688412, "lr_stage": "cosine_decay", "mfu": 60.810938, "stage_name": "permutation_cycle", "status": "training", "step": 1807, "tokens_per_sec_per_gpu": 105861.520517, "vram_gb": 20.709735} {"adam_lr": 2.7811814881259503e-06, "dt_ms": 155.224085, "global_tokens_per_sec": 105550.630338, "grad_norm": 21.875, "loss": 2.476103, "lr_stage": "cosine_decay", "mfu": 60.63235, "stage_name": "permutation_cycle", "status": "training", "step": 1808, "tokens_per_sec_per_gpu": 105550.630338, "vram_gb": 20.709735} {"adam_lr": 2.752554392123463e-06, "dt_ms": 152.919292, "global_tokens_per_sec": 107141.484488, "grad_norm": 36.5, "loss": 1.889646, "lr_stage": "cosine_decay", "mfu": 61.546198, "stage_name": "permutation_cycle", "status": "training", "step": 1809, "tokens_per_sec_per_gpu": 107141.484488, "vram_gb": 20.709735} {"adam_lr": 2.724071220034158e-06, "dt_ms": 153.50914, "global_tokens_per_sec": 106729.801225, "grad_norm": 31.625, "loss": 2.550395, "lr_stage": "cosine_decay", "mfu": 61.309711, "stage_name": "permutation_cycle", "status": "training", "step": 1810, "tokens_per_sec_per_gpu": 106729.801225, "vram_gb": 20.709735} {"adam_lr": 2.695732058622735e-06, "dt_ms": 157.356501, "global_tokens_per_sec": 104120.261539, "grad_norm": 31.875, "loss": 2.214602, "lr_stage": "cosine_decay", "mfu": 59.810691, "stage_name": "permutation_cycle", "status": "training", "step": 1811, "tokens_per_sec_per_gpu": 104120.261539, "vram_gb": 20.709735} {"adam_lr": 2.667536994215186e-06, "dt_ms": 157.124519, "global_tokens_per_sec": 104273.986441, "grad_norm": 39.25, "loss": 2.409283, "lr_stage": "cosine_decay", "mfu": 59.898997, "stage_name": "permutation_cycle", "status": "training", "step": 1812, "tokens_per_sec_per_gpu": 104273.986441, "vram_gb": 20.709735} {"adam_lr": 2.6394861126985636e-06, "dt_ms": 154.576063, "global_tokens_per_sec": 105993.125103, "grad_norm": 29.875, "loss": 2.362523, "lr_stage": "cosine_decay", "mfu": 60.886536, "stage_name": "permutation_cycle", "status": "training", "step": 1813, "tokens_per_sec_per_gpu": 105993.125103, "vram_gb": 20.709735} {"adam_lr": 2.611579499520722e-06, "dt_ms": 155.976057, "global_tokens_per_sec": 105041.762881, "grad_norm": 24.875, "loss": 2.155424, "lr_stage": "cosine_decay", "mfu": 60.340037, "stage_name": "permutation_cycle", "status": "training", "step": 1814, "tokens_per_sec_per_gpu": 105041.762881, "vram_gb": 20.709735} {"adam_lr": 2.583817239690034e-06, "dt_ms": 155.303001, "global_tokens_per_sec": 105496.995241, "grad_norm": 30.625, "loss": 1.964319, "lr_stage": "cosine_decay", "mfu": 60.60154, "stage_name": "permutation_cycle", "status": "training", "step": 1815, "tokens_per_sec_per_gpu": 105496.995241, "vram_gb": 20.709735} {"adam_lr": 2.5561994177751737e-06, "dt_ms": 154.522419, "global_tokens_per_sec": 106029.921798, "grad_norm": 35.0, "loss": 1.832562, "lr_stage": "cosine_decay", "mfu": 60.907674, "stage_name": "permutation_cycle", "status": "training", "step": 1816, "tokens_per_sec_per_gpu": 106029.921798, "vram_gb": 20.709735} {"adam_lr": 2.528726117904812e-06, "dt_ms": 156.215668, "global_tokens_per_sec": 104880.645064, "grad_norm": 34.5, "loss": 3.023512, "lr_stage": "cosine_decay", "mfu": 60.247485, "stage_name": "permutation_cycle", "status": "training", "step": 1817, "tokens_per_sec_per_gpu": 104880.645064, "vram_gb": 20.709735} {"adam_lr": 2.501397423767382e-06, "dt_ms": 155.683994, "global_tokens_per_sec": 105238.820949, "grad_norm": 44.0, "loss": 3.117674, "lr_stage": "cosine_decay", "mfu": 60.453235, "stage_name": "permutation_cycle", "status": "training", "step": 1818, "tokens_per_sec_per_gpu": 105238.820949, "vram_gb": 20.709735} {"adam_lr": 2.4742134186108155e-06, "dt_ms": 154.841185, "global_tokens_per_sec": 105811.642042, "grad_norm": 26.75, "loss": 2.120152, "lr_stage": "cosine_decay", "mfu": 60.782285, "stage_name": "permutation_cycle", "status": "training", "step": 1819, "tokens_per_sec_per_gpu": 105811.642042, "vram_gb": 20.709735} {"adam_lr": 2.4471741852423237e-06, "dt_ms": 155.184746, "global_tokens_per_sec": 105577.387241, "grad_norm": 31.625, "loss": 2.090284, "lr_stage": "cosine_decay", "mfu": 60.64772, "stage_name": "permutation_cycle", "status": "training", "step": 1820, "tokens_per_sec_per_gpu": 105577.387241, "vram_gb": 20.709735} {"adam_lr": 2.4202798060280916e-06, "dt_ms": 155.015707, "global_tokens_per_sec": 105692.515393, "grad_norm": 32.0, "loss": 2.342778, "lr_stage": "cosine_decay", "mfu": 60.713854, "stage_name": "permutation_cycle", "status": "training", "step": 1821, "tokens_per_sec_per_gpu": 105692.515393, "vram_gb": 20.709735} {"adam_lr": 2.3935303628930707e-06, "dt_ms": 155.899286, "global_tokens_per_sec": 105093.48947, "grad_norm": 27.25, "loss": 2.487164, "lr_stage": "cosine_decay", "mfu": 60.369751, "stage_name": "permutation_cycle", "status": "training", "step": 1822, "tokens_per_sec_per_gpu": 105093.48947, "vram_gb": 20.709735} {"adam_lr": 2.366925937320691e-06, "dt_ms": 153.669119, "global_tokens_per_sec": 106618.689033, "grad_norm": 35.5, "loss": 2.798713, "lr_stage": "cosine_decay", "mfu": 61.245884, "stage_name": "permutation_cycle", "status": "training", "step": 1823, "tokens_per_sec_per_gpu": 106618.689033, "vram_gb": 20.709735} {"adam_lr": 2.340466610352654e-06, "dt_ms": 154.94132, "global_tokens_per_sec": 105743.257871, "grad_norm": 23.5, "loss": 2.407539, "lr_stage": "cosine_decay", "mfu": 60.743003, "stage_name": "permutation_cycle", "status": "training", "step": 1824, "tokens_per_sec_per_gpu": 105743.257871, "vram_gb": 20.709735} {"adam_lr": 2.314152462588659e-06, "dt_ms": 155.774593, "global_tokens_per_sec": 105177.613674, "grad_norm": 33.25, "loss": 2.654753, "lr_stage": "cosine_decay", "mfu": 60.418075, "stage_name": "permutation_cycle", "status": "training", "step": 1825, "tokens_per_sec_per_gpu": 105177.613674, "vram_gb": 20.709735} {"adam_lr": 2.2879835741861586e-06, "dt_ms": 155.266285, "global_tokens_per_sec": 105521.942552, "grad_norm": 25.5, "loss": 2.65278, "lr_stage": "cosine_decay", "mfu": 60.615871, "stage_name": "permutation_cycle", "status": "training", "step": 1826, "tokens_per_sec_per_gpu": 105521.942552, "vram_gb": 20.709735} {"adam_lr": 2.2619600248601327e-06, "dt_ms": 155.55048, "global_tokens_per_sec": 105329.151101, "grad_norm": 34.5, "loss": 2.42535, "lr_stage": "cosine_decay", "mfu": 60.505124, "stage_name": "permutation_cycle", "status": "training", "step": 1827, "tokens_per_sec_per_gpu": 105329.151101, "vram_gb": 20.709735} {"adam_lr": 2.2360818938828187e-06, "dt_ms": 155.404568, "global_tokens_per_sec": 105428.046553, "grad_norm": 14.25, "loss": 1.78538, "lr_stage": "cosine_decay", "mfu": 60.561933, "stage_name": "permutation_cycle", "status": "training", "step": 1828, "tokens_per_sec_per_gpu": 105428.046553, "vram_gb": 20.709735} {"adam_lr": 2.210349260083494e-06, "dt_ms": 154.973745, "global_tokens_per_sec": 105721.133366, "grad_norm": 29.75, "loss": 2.619998, "lr_stage": "cosine_decay", "mfu": 60.730294, "stage_name": "permutation_cycle", "status": "training", "step": 1829, "tokens_per_sec_per_gpu": 105721.133366, "vram_gb": 20.709735} {"adam_lr": 2.1847622018482283e-06, "dt_ms": 155.607224, "global_tokens_per_sec": 105290.741846, "grad_norm": 40.75, "loss": 2.430349, "lr_stage": "cosine_decay", "mfu": 60.48306, "stage_name": "permutation_cycle", "status": "training", "step": 1830, "tokens_per_sec_per_gpu": 105290.741846, "vram_gb": 20.709735} {"adam_lr": 2.1593207971196296e-06, "dt_ms": 154.765368, "global_tokens_per_sec": 105863.477494, "grad_norm": 24.125, "loss": 2.039848, "lr_stage": "cosine_decay", "mfu": 60.812062, "stage_name": "permutation_cycle", "status": "training", "step": 1831, "tokens_per_sec_per_gpu": 105863.477494, "vram_gb": 20.709735} {"adam_lr": 2.134025123396638e-06, "dt_ms": 154.658794, "global_tokens_per_sec": 105936.426462, "grad_norm": 34.0, "loss": 1.952053, "lr_stage": "cosine_decay", "mfu": 60.853966, "stage_name": "permutation_cycle", "status": "training", "step": 1832, "tokens_per_sec_per_gpu": 105936.426462, "vram_gb": 20.709735} {"adam_lr": 2.108875257734261e-06, "dt_ms": 154.210091, "global_tokens_per_sec": 106244.668765, "grad_norm": 42.75, "loss": 2.287819, "lr_stage": "cosine_decay", "mfu": 61.031033, "stage_name": "permutation_cycle", "status": "training", "step": 1833, "tokens_per_sec_per_gpu": 106244.668765, "vram_gb": 20.709735} {"adam_lr": 2.0838712767433375e-06, "dt_ms": 154.604197, "global_tokens_per_sec": 105973.837488, "grad_norm": 20.875, "loss": 1.162811, "lr_stage": "cosine_decay", "mfu": 60.875457, "stage_name": "permutation_cycle", "status": "training", "step": 1834, "tokens_per_sec_per_gpu": 105973.837488, "vram_gb": 20.709735} {"adam_lr": 2.0590132565903476e-06, "dt_ms": 154.641867, "global_tokens_per_sec": 105948.022688, "grad_norm": 35.0, "loss": 2.148227, "lr_stage": "cosine_decay", "mfu": 60.860628, "stage_name": "permutation_cycle", "status": "training", "step": 1835, "tokens_per_sec_per_gpu": 105948.022688, "vram_gb": 20.709735} {"adam_lr": 2.0343012729971243e-06, "dt_ms": 154.982328, "global_tokens_per_sec": 105715.278429, "grad_norm": 31.25, "loss": 2.911725, "lr_stage": "cosine_decay", "mfu": 60.72693, "stage_name": "permutation_cycle", "status": "training", "step": 1836, "tokens_per_sec_per_gpu": 105715.278429, "vram_gb": 20.709735} {"adam_lr": 2.0097354012406535e-06, "dt_ms": 154.189587, "global_tokens_per_sec": 106258.797089, "grad_norm": 41.25, "loss": 2.42481, "lr_stage": "cosine_decay", "mfu": 61.039149, "stage_name": "permutation_cycle", "status": "training", "step": 1837, "tokens_per_sec_per_gpu": 106258.797089, "vram_gb": 20.709735} {"adam_lr": 1.985315716152847e-06, "dt_ms": 155.030012, "global_tokens_per_sec": 105682.762807, "grad_norm": 27.125, "loss": 2.091553, "lr_stage": "cosine_decay", "mfu": 60.708252, "stage_name": "permutation_cycle", "status": "training", "step": 1838, "tokens_per_sec_per_gpu": 105682.762807, "vram_gb": 20.709735} {"adam_lr": 1.961042292120291e-06, "dt_ms": 155.356884, "global_tokens_per_sec": 105460.405602, "grad_norm": 26.75, "loss": 2.201351, "lr_stage": "cosine_decay", "mfu": 60.580522, "stage_name": "permutation_cycle", "status": "training", "step": 1839, "tokens_per_sec_per_gpu": 105460.405602, "vram_gb": 20.709735} {"adam_lr": 1.9369152030840556e-06, "dt_ms": 154.527903, "global_tokens_per_sec": 106026.159185, "grad_norm": 29.25, "loss": 2.023393, "lr_stage": "cosine_decay", "mfu": 60.905512, "stage_name": "permutation_cycle", "status": "training", "step": 1840, "tokens_per_sec_per_gpu": 106026.159185, "vram_gb": 20.709735} {"adam_lr": 1.9129345225394334e-06, "dt_ms": 154.808998, "global_tokens_per_sec": 105833.641457, "grad_norm": 42.0, "loss": 2.298156, "lr_stage": "cosine_decay", "mfu": 60.794923, "stage_name": "permutation_cycle", "status": "training", "step": 1841, "tokens_per_sec_per_gpu": 105833.641457, "vram_gb": 20.709735} {"adam_lr": 1.8891003235357308e-06, "dt_ms": 155.731201, "global_tokens_per_sec": 105206.919851, "grad_norm": 25.125, "loss": 2.081267, "lr_stage": "cosine_decay", "mfu": 60.43491, "stage_name": "permutation_cycle", "status": "training", "step": 1842, "tokens_per_sec_per_gpu": 105206.919851, "vram_gb": 20.709735} {"adam_lr": 1.8654126786760596e-06, "dt_ms": 154.789925, "global_tokens_per_sec": 105846.682464, "grad_norm": 40.75, "loss": 2.839797, "lr_stage": "cosine_decay", "mfu": 60.802414, "stage_name": "permutation_cycle", "status": "training", "step": 1843, "tokens_per_sec_per_gpu": 105846.682464, "vram_gb": 20.709735} {"adam_lr": 1.841871660117095e-06, "dt_ms": 155.317783, "global_tokens_per_sec": 105486.954848, "grad_norm": 29.125, "loss": 2.255255, "lr_stage": "cosine_decay", "mfu": 60.595773, "stage_name": "permutation_cycle", "status": "training", "step": 1844, "tokens_per_sec_per_gpu": 105486.954848, "vram_gb": 20.709735} {"adam_lr": 1.8184773395688526e-06, "dt_ms": 154.927969, "global_tokens_per_sec": 105752.370653, "grad_norm": 32.75, "loss": 2.660562, "lr_stage": "cosine_decay", "mfu": 60.748238, "stage_name": "permutation_cycle", "status": "training", "step": 1845, "tokens_per_sec_per_gpu": 105752.370653, "vram_gb": 20.709735} {"adam_lr": 1.7952297882945003e-06, "dt_ms": 154.761076, "global_tokens_per_sec": 105866.413095, "grad_norm": 42.75, "loss": 2.429108, "lr_stage": "cosine_decay", "mfu": 60.813748, "stage_name": "permutation_cycle", "status": "training", "step": 1846, "tokens_per_sec_per_gpu": 105866.413095, "vram_gb": 20.709735} {"adam_lr": 1.7721290771100961e-06, "dt_ms": 154.65188, "global_tokens_per_sec": 105941.162642, "grad_norm": 33.25, "loss": 2.67117, "lr_stage": "cosine_decay", "mfu": 60.856687, "stage_name": "permutation_cycle", "status": "training", "step": 1847, "tokens_per_sec_per_gpu": 105941.162642, "vram_gb": 20.709735} {"adam_lr": 1.7491752763844293e-06, "dt_ms": 155.558825, "global_tokens_per_sec": 105323.500923, "grad_norm": 38.75, "loss": 2.330086, "lr_stage": "cosine_decay", "mfu": 60.501878, "stage_name": "permutation_cycle", "status": "training", "step": 1848, "tokens_per_sec_per_gpu": 105323.500923, "vram_gb": 20.709735} {"adam_lr": 1.7263684560387516e-06, "dt_ms": 156.119347, "global_tokens_per_sec": 104945.353378, "grad_norm": 29.75, "loss": 2.570929, "lr_stage": "cosine_decay", "mfu": 60.284656, "stage_name": "permutation_cycle", "status": "training", "step": 1849, "tokens_per_sec_per_gpu": 104945.353378, "vram_gb": 20.709735} {"adam_lr": 1.70370868554659e-06, "dt_ms": 153.407097, "global_tokens_per_sec": 106800.795628, "grad_norm": 13.0, "loss": 0.881349, "lr_stage": "cosine_decay", "mfu": 61.350493, "stage_name": "permutation_cycle", "status": "training", "step": 1850, "tokens_per_sec_per_gpu": 106800.795628, "vram_gb": 20.709735} {"adam_lr": 1.6811960339335298e-06, "dt_ms": 155.265808, "global_tokens_per_sec": 105522.266621, "grad_norm": 35.25, "loss": 2.675931, "lr_stage": "cosine_decay", "mfu": 60.616057, "stage_name": "permutation_cycle", "status": "training", "step": 1851, "tokens_per_sec_per_gpu": 105522.266621, "vram_gb": 20.709735} {"adam_lr": 1.6588305697770313e-06, "dt_ms": 155.427694, "global_tokens_per_sec": 105412.359564, "grad_norm": 28.875, "loss": 1.74768, "lr_stage": "cosine_decay", "mfu": 60.552922, "stage_name": "permutation_cycle", "status": "training", "step": 1852, "tokens_per_sec_per_gpu": 105412.359564, "vram_gb": 20.709735} {"adam_lr": 1.6366123612061636e-06, "dt_ms": 154.845476, "global_tokens_per_sec": 105808.709478, "grad_norm": 26.875, "loss": 2.406893, "lr_stage": "cosine_decay", "mfu": 60.780601, "stage_name": "permutation_cycle", "status": "training", "step": 1853, "tokens_per_sec_per_gpu": 105808.709478, "vram_gb": 20.709735} {"adam_lr": 1.6145414759014431e-06, "dt_ms": 155.641556, "global_tokens_per_sec": 105267.516231, "grad_norm": 43.75, "loss": 2.044965, "lr_stage": "cosine_decay", "mfu": 60.469719, "stage_name": "permutation_cycle", "status": "training", "step": 1854, "tokens_per_sec_per_gpu": 105267.516231, "vram_gb": 20.709735} {"adam_lr": 1.5926179810946184e-06, "dt_ms": 155.26104, "global_tokens_per_sec": 105525.507417, "grad_norm": 28.75, "loss": 2.068034, "lr_stage": "cosine_decay", "mfu": 60.617919, "stage_name": "permutation_cycle", "status": "training", "step": 1855, "tokens_per_sec_per_gpu": 105525.507417, "vram_gb": 20.709735} {"adam_lr": 1.5708419435684462e-06, "dt_ms": 154.748917, "global_tokens_per_sec": 105874.731515, "grad_norm": 29.25, "loss": 1.958089, "lr_stage": "cosine_decay", "mfu": 60.818526, "stage_name": "permutation_cycle", "status": "training", "step": 1856, "tokens_per_sec_per_gpu": 105874.731515, "vram_gb": 20.709735} {"adam_lr": 1.5492134296565264e-06, "dt_ms": 156.026125, "global_tokens_per_sec": 105008.055573, "grad_norm": 33.75, "loss": 3.038473, "lr_stage": "cosine_decay", "mfu": 60.320674, "stage_name": "permutation_cycle", "status": "training", "step": 1857, "tokens_per_sec_per_gpu": 105008.055573, "vram_gb": 20.709735} {"adam_lr": 1.5277325052430568e-06, "dt_ms": 156.241655, "global_tokens_per_sec": 104863.200299, "grad_norm": 25.125, "loss": 2.139809, "lr_stage": "cosine_decay", "mfu": 60.237464, "stage_name": "permutation_cycle", "status": "training", "step": 1858, "tokens_per_sec_per_gpu": 104863.200299, "vram_gb": 20.709735} {"adam_lr": 1.506399235762662e-06, "dt_ms": 155.31826, "global_tokens_per_sec": 105486.630997, "grad_norm": 39.25, "loss": 2.311156, "lr_stage": "cosine_decay", "mfu": 60.595587, "stage_name": "permutation_cycle", "status": "training", "step": 1859, "tokens_per_sec_per_gpu": 105486.630997, "vram_gb": 20.709735} {"adam_lr": 1.4852136862001764e-06, "dt_ms": 158.667564, "global_tokens_per_sec": 103259.919964, "grad_norm": 39.5, "loss": 2.210461, "lr_stage": "cosine_decay", "mfu": 59.316478, "stage_name": "permutation_cycle", "status": "training", "step": 1860, "tokens_per_sec_per_gpu": 103259.919964, "vram_gb": 20.709735} {"adam_lr": 1.4641759210904605e-06, "dt_ms": 155.882835, "global_tokens_per_sec": 105104.580368, "grad_norm": 21.25, "loss": 2.328857, "lr_stage": "cosine_decay", "mfu": 60.376122, "stage_name": "permutation_cycle", "status": "training", "step": 1861, "tokens_per_sec_per_gpu": 105104.580368, "vram_gb": 20.709735} {"adam_lr": 1.4432860045182017e-06, "dt_ms": 154.545307, "global_tokens_per_sec": 106014.21875, "grad_norm": 21.75, "loss": 2.450574, "lr_stage": "cosine_decay", "mfu": 60.898653, "stage_name": "permutation_cycle", "status": "training", "step": 1862, "tokens_per_sec_per_gpu": 106014.21875, "vram_gb": 20.709735} {"adam_lr": 1.422544000117698e-06, "dt_ms": 153.687477, "global_tokens_per_sec": 106605.953249, "grad_norm": 29.125, "loss": 2.157643, "lr_stage": "cosine_decay", "mfu": 61.238568, "stage_name": "permutation_cycle", "status": "training", "step": 1863, "tokens_per_sec_per_gpu": 106605.953249, "vram_gb": 20.709735} {"adam_lr": 1.4019499710726913e-06, "dt_ms": 154.215097, "global_tokens_per_sec": 106241.219396, "grad_norm": 38.75, "loss": 2.508349, "lr_stage": "cosine_decay", "mfu": 61.029051, "stage_name": "permutation_cycle", "status": "training", "step": 1864, "tokens_per_sec_per_gpu": 106241.219396, "vram_gb": 20.709735} {"adam_lr": 1.3815039801161721e-06, "dt_ms": 154.902697, "global_tokens_per_sec": 105769.624149, "grad_norm": 34.25, "loss": 1.652957, "lr_stage": "cosine_decay", "mfu": 60.758149, "stage_name": "permutation_cycle", "status": "training", "step": 1865, "tokens_per_sec_per_gpu": 105769.624149, "vram_gb": 20.709735} {"adam_lr": 1.3612060895301759e-06, "dt_ms": 155.606985, "global_tokens_per_sec": 105290.903171, "grad_norm": 25.625, "loss": 2.527452, "lr_stage": "cosine_decay", "mfu": 60.483153, "stage_name": "permutation_cycle", "status": "training", "step": 1866, "tokens_per_sec_per_gpu": 105290.903171, "vram_gb": 20.709735} {"adam_lr": 1.3410563611455928e-06, "dt_ms": 155.8671, "global_tokens_per_sec": 105115.191243, "grad_norm": 31.25, "loss": 2.815732, "lr_stage": "cosine_decay", "mfu": 60.382217, "stage_name": "permutation_cycle", "status": "training", "step": 1867, "tokens_per_sec_per_gpu": 105115.191243, "vram_gb": 20.709735} {"adam_lr": 1.3210548563419856e-06, "dt_ms": 154.887915, "global_tokens_per_sec": 105779.718426, "grad_norm": 40.5, "loss": 2.369132, "lr_stage": "cosine_decay", "mfu": 60.763947, "stage_name": "permutation_cycle", "status": "training", "step": 1868, "tokens_per_sec_per_gpu": 105779.718426, "vram_gb": 20.709735} {"adam_lr": 1.3012016360474221e-06, "dt_ms": 155.291557, "global_tokens_per_sec": 105504.769761, "grad_norm": 27.25, "loss": 2.181842, "lr_stage": "cosine_decay", "mfu": 60.606006, "stage_name": "permutation_cycle", "status": "training", "step": 1869, "tokens_per_sec_per_gpu": 105504.769761, "vram_gb": 20.709735} {"adam_lr": 1.2814967607382432e-06, "dt_ms": 156.774044, "global_tokens_per_sec": 104507.095551, "grad_norm": 26.0, "loss": 2.752015, "lr_stage": "cosine_decay", "mfu": 60.032904, "stage_name": "permutation_cycle", "status": "training", "step": 1870, "tokens_per_sec_per_gpu": 104507.095551, "vram_gb": 20.709735} {"adam_lr": 1.2619402904389121e-06, "dt_ms": 155.025721, "global_tokens_per_sec": 105685.688394, "grad_norm": 36.25, "loss": 2.451506, "lr_stage": "cosine_decay", "mfu": 60.709933, "stage_name": "permutation_cycle", "status": "training", "step": 1871, "tokens_per_sec_per_gpu": 105685.688394, "vram_gb": 20.709735} {"adam_lr": 1.2425322847218368e-06, "dt_ms": 155.166626, "global_tokens_per_sec": 105589.716196, "grad_norm": 26.25, "loss": 1.569629, "lr_stage": "cosine_decay", "mfu": 60.654803, "stage_name": "permutation_cycle", "status": "training", "step": 1872, "tokens_per_sec_per_gpu": 105589.716196, "vram_gb": 20.709735} {"adam_lr": 1.2232728027071705e-06, "dt_ms": 154.5434, "global_tokens_per_sec": 106015.52716, "grad_norm": 31.0, "loss": 2.618757, "lr_stage": "cosine_decay", "mfu": 60.899405, "stage_name": "permutation_cycle", "status": "training", "step": 1873, "tokens_per_sec_per_gpu": 106015.52716, "vram_gb": 20.709735} {"adam_lr": 1.2041619030626284e-06, "dt_ms": 154.812574, "global_tokens_per_sec": 105831.196625, "grad_norm": 28.125, "loss": 2.283945, "lr_stage": "cosine_decay", "mfu": 60.793518, "stage_name": "permutation_cycle", "status": "training", "step": 1874, "tokens_per_sec_per_gpu": 105831.196625, "vram_gb": 20.709735} {"adam_lr": 1.1851996440033319e-06, "dt_ms": 156.2047, "global_tokens_per_sec": 104888.008816, "grad_norm": 24.5, "loss": 1.677157, "lr_stage": "cosine_decay", "mfu": 60.251715, "stage_name": "permutation_cycle", "status": "training", "step": 1875, "tokens_per_sec_per_gpu": 104888.008816, "vram_gb": 20.709735} {"adam_lr": 1.166386083291604e-06, "dt_ms": 155.892134, "global_tokens_per_sec": 105098.311312, "grad_norm": 27.25, "loss": 2.968796, "lr_stage": "cosine_decay", "mfu": 60.372521, "stage_name": "permutation_cycle", "status": "training", "step": 1876, "tokens_per_sec_per_gpu": 105098.311312, "vram_gb": 20.709735} {"adam_lr": 1.1477212782368185e-06, "dt_ms": 155.291319, "global_tokens_per_sec": 105504.931742, "grad_norm": 22.625, "loss": 2.310344, "lr_stage": "cosine_decay", "mfu": 60.606099, "stage_name": "permutation_cycle", "status": "training", "step": 1877, "tokens_per_sec_per_gpu": 105504.931742, "vram_gb": 20.709735} {"adam_lr": 1.1292052856952062e-06, "dt_ms": 155.770302, "global_tokens_per_sec": 105180.51136, "grad_norm": 36.0, "loss": 2.664565, "lr_stage": "cosine_decay", "mfu": 60.41974, "stage_name": "permutation_cycle", "status": "training", "step": 1878, "tokens_per_sec_per_gpu": 105180.51136, "vram_gb": 20.709735} {"adam_lr": 1.1108381620696883e-06, "dt_ms": 156.737566, "global_tokens_per_sec": 104531.417826, "grad_norm": 25.75, "loss": 2.146563, "lr_stage": "cosine_decay", "mfu": 60.046875, "stage_name": "permutation_cycle", "status": "training", "step": 1879, "tokens_per_sec_per_gpu": 104531.417826, "vram_gb": 20.709735} {"adam_lr": 1.0926199633097157e-06, "dt_ms": 154.13785, "global_tokens_per_sec": 106294.463173, "grad_norm": 20.5, "loss": 1.644798, "lr_stage": "cosine_decay", "mfu": 61.059636, "stage_name": "permutation_cycle", "status": "training", "step": 1880, "tokens_per_sec_per_gpu": 106294.463173, "vram_gb": 20.709735} {"adam_lr": 1.0745507449110791e-06, "dt_ms": 156.048298, "global_tokens_per_sec": 104993.134961, "grad_norm": 39.0, "loss": 2.687409, "lr_stage": "cosine_decay", "mfu": 60.312103, "stage_name": "permutation_cycle", "status": "training", "step": 1881, "tokens_per_sec_per_gpu": 104993.134961, "vram_gb": 20.709735} {"adam_lr": 1.0566305619157502e-06, "dt_ms": 156.113863, "global_tokens_per_sec": 104949.039669, "grad_norm": 23.875, "loss": 2.867899, "lr_stage": "cosine_decay", "mfu": 60.286773, "stage_name": "permutation_cycle", "status": "training", "step": 1882, "tokens_per_sec_per_gpu": 104949.039669, "vram_gb": 20.709735} {"adam_lr": 1.0388594689117071e-06, "dt_ms": 156.784058, "global_tokens_per_sec": 104500.420827, "grad_norm": 35.0, "loss": 2.897217, "lr_stage": "cosine_decay", "mfu": 60.029069, "stage_name": "permutation_cycle", "status": "training", "step": 1883, "tokens_per_sec_per_gpu": 104500.420827, "vram_gb": 20.709735} {"adam_lr": 1.0212375200327973e-06, "dt_ms": 157.480717, "global_tokens_per_sec": 104038.134591, "grad_norm": 25.375, "loss": 2.751252, "lr_stage": "cosine_decay", "mfu": 59.763514, "stage_name": "permutation_cycle", "status": "training", "step": 1884, "tokens_per_sec_per_gpu": 104038.134591, "vram_gb": 20.709735} {"adam_lr": 1.0037647689585206e-06, "dt_ms": 157.513142, "global_tokens_per_sec": 104016.717781, "grad_norm": 33.5, "loss": 2.825094, "lr_stage": "cosine_decay", "mfu": 59.751212, "stage_name": "permutation_cycle", "status": "training", "step": 1885, "tokens_per_sec_per_gpu": 104016.717781, "vram_gb": 20.709735} {"adam_lr": 9.864412689139123e-07, "dt_ms": 157.068968, "global_tokens_per_sec": 104310.86565, "grad_norm": 32.5, "loss": 2.095372, "lr_stage": "cosine_decay", "mfu": 59.920182, "stage_name": "permutation_cycle", "status": "training", "step": 1886, "tokens_per_sec_per_gpu": 104310.86565, "vram_gb": 20.709735} {"adam_lr": 9.692670726693498e-07, "dt_ms": 157.146692, "global_tokens_per_sec": 104259.273693, "grad_norm": 32.5, "loss": 1.60788, "lr_stage": "cosine_decay", "mfu": 59.890545, "stage_name": "permutation_cycle", "status": "training", "step": 1887, "tokens_per_sec_per_gpu": 104259.273693, "vram_gb": 20.709735} {"adam_lr": 9.522422325404235e-07, "dt_ms": 157.895803, "global_tokens_per_sec": 103764.632383, "grad_norm": 33.75, "loss": 2.556188, "lr_stage": "cosine_decay", "mfu": 59.606404, "stage_name": "permutation_cycle", "status": "training", "step": 1888, "tokens_per_sec_per_gpu": 103764.632383, "vram_gb": 20.709735} {"adam_lr": 9.353668003877437e-07, "dt_ms": 157.153606, "global_tokens_per_sec": 104254.686697, "grad_norm": 28.875, "loss": 2.253557, "lr_stage": "cosine_decay", "mfu": 59.88791, "stage_name": "permutation_cycle", "status": "training", "step": 1889, "tokens_per_sec_per_gpu": 104254.686697, "vram_gb": 20.709735} {"adam_lr": 9.186408276168013e-07, "dt_ms": 158.881426, "global_tokens_per_sec": 103120.927519, "grad_norm": 27.5, "loss": 2.020522, "lr_stage": "cosine_decay", "mfu": 59.236635, "stage_name": "permutation_cycle", "status": "training", "step": 1890, "tokens_per_sec_per_gpu": 103120.927519, "vram_gb": 20.709735} {"adam_lr": 9.020643651778182e-07, "dt_ms": 157.025814, "global_tokens_per_sec": 104339.532315, "grad_norm": 32.0, "loss": 2.568133, "lr_stage": "cosine_decay", "mfu": 59.936649, "stage_name": "permutation_cycle", "status": "training", "step": 1891, "tokens_per_sec_per_gpu": 104339.532315, "vram_gb": 20.709735} {"adam_lr": 8.856374635655695e-07, "dt_ms": 159.358025, "global_tokens_per_sec": 102812.519429, "grad_norm": 27.875, "loss": 1.727524, "lr_stage": "cosine_decay", "mfu": 59.059474, "stage_name": "permutation_cycle", "status": "training", "step": 1892, "tokens_per_sec_per_gpu": 102812.519429, "vram_gb": 20.709735} {"adam_lr": 8.693601728192391e-07, "dt_ms": 158.939362, "global_tokens_per_sec": 103083.338438, "grad_norm": 16.75, "loss": 2.284252, "lr_stage": "cosine_decay", "mfu": 59.215043, "stage_name": "permutation_cycle", "status": "training", "step": 1893, "tokens_per_sec_per_gpu": 103083.338438, "vram_gb": 20.709735} {"adam_lr": 8.53232542522292e-07, "dt_ms": 159.667253, "global_tokens_per_sec": 102613.401568, "grad_norm": 32.0, "loss": 2.669583, "lr_stage": "cosine_decay", "mfu": 58.945093, "stage_name": "permutation_cycle", "status": "training", "step": 1894, "tokens_per_sec_per_gpu": 102613.401568, "vram_gb": 20.709735} {"adam_lr": 8.372546218022747e-07, "dt_ms": 156.271935, "global_tokens_per_sec": 104842.882066, "grad_norm": 28.625, "loss": 1.823738, "lr_stage": "cosine_decay", "mfu": 60.225792, "stage_name": "permutation_cycle", "status": "training", "step": 1895, "tokens_per_sec_per_gpu": 104842.882066, "vram_gb": 20.709735} {"adam_lr": 8.214264593307098e-07, "dt_ms": 156.998158, "global_tokens_per_sec": 104357.912607, "grad_norm": 26.25, "loss": 2.798975, "lr_stage": "cosine_decay", "mfu": 59.947207, "stage_name": "permutation_cycle", "status": "training", "step": 1896, "tokens_per_sec_per_gpu": 104357.912607, "vram_gb": 20.709735} {"adam_lr": 8.057481033229175e-07, "dt_ms": 154.986382, "global_tokens_per_sec": 105712.513823, "grad_norm": 31.125, "loss": 2.108053, "lr_stage": "cosine_decay", "mfu": 60.725342, "stage_name": "permutation_cycle", "status": "training", "step": 1897, "tokens_per_sec_per_gpu": 105712.513823, "vram_gb": 20.709735} {"adam_lr": 7.90219601537906e-07, "dt_ms": 154.369354, "global_tokens_per_sec": 106135.055626, "grad_norm": 19.75, "loss": 2.320932, "lr_stage": "cosine_decay", "mfu": 60.968067, "stage_name": "permutation_cycle", "status": "training", "step": 1898, "tokens_per_sec_per_gpu": 106135.055626, "vram_gb": 20.709735} {"adam_lr": 7.748410012781704e-07, "dt_ms": 154.089212, "global_tokens_per_sec": 106328.014421, "grad_norm": 24.0, "loss": 2.560853, "lr_stage": "cosine_decay", "mfu": 61.07891, "stage_name": "permutation_cycle", "status": "training", "step": 1899, "tokens_per_sec_per_gpu": 106328.014421, "vram_gb": 20.709735} {"adam_lr": 7.596123493895991e-07, "dt_ms": 152.795792, "global_tokens_per_sec": 107228.084135, "grad_norm": 33.0, "loss": 1.613573, "lr_stage": "cosine_decay", "mfu": 61.595944, "stage_name": "permutation_cycle", "status": "training", "step": 1900, "tokens_per_sec_per_gpu": 107228.084135, "vram_gb": 20.709735} {"adam_lr": 7.445336922613067e-07, "dt_ms": 155.341387, "global_tokens_per_sec": 105470.92657, "grad_norm": 24.5, "loss": 3.100365, "lr_stage": "cosine_decay", "mfu": 60.586565, "stage_name": "permutation_cycle", "status": "training", "step": 1901, "tokens_per_sec_per_gpu": 105470.92657, "vram_gb": 20.709735} {"adam_lr": 7.296050758254957e-07, "dt_ms": 155.957937, "global_tokens_per_sec": 105053.96705, "grad_norm": 35.5, "loss": 2.97193, "lr_stage": "cosine_decay", "mfu": 60.347048, "stage_name": "permutation_cycle", "status": "training", "step": 1902, "tokens_per_sec_per_gpu": 105053.96705, "vram_gb": 20.709735} {"adam_lr": 7.148265455573233e-07, "dt_ms": 156.10528, "global_tokens_per_sec": 104954.810037, "grad_norm": 30.0, "loss": 2.026497, "lr_stage": "cosine_decay", "mfu": 60.290088, "stage_name": "permutation_cycle", "status": "training", "step": 1903, "tokens_per_sec_per_gpu": 104954.810037, "vram_gb": 20.709735} {"adam_lr": 7.001981464747565e-07, "dt_ms": 155.016661, "global_tokens_per_sec": 105691.865165, "grad_norm": 20.125, "loss": 2.362324, "lr_stage": "cosine_decay", "mfu": 60.713481, "stage_name": "permutation_cycle", "status": "training", "step": 1904, "tokens_per_sec_per_gpu": 105691.865165, "vram_gb": 20.709735} {"adam_lr": 6.857199231384282e-07, "dt_ms": 154.02174, "global_tokens_per_sec": 106374.59364, "grad_norm": 25.125, "loss": 1.64392, "lr_stage": "cosine_decay", "mfu": 61.105667, "stage_name": "permutation_cycle", "status": "training", "step": 1905, "tokens_per_sec_per_gpu": 106374.59364, "vram_gb": 20.709735} {"adam_lr": 6.713919196515317e-07, "dt_ms": 154.066086, "global_tokens_per_sec": 106343.975141, "grad_norm": 37.5, "loss": 2.143725, "lr_stage": "cosine_decay", "mfu": 61.088078, "stage_name": "permutation_cycle", "status": "training", "step": 1906, "tokens_per_sec_per_gpu": 106343.975141, "vram_gb": 20.709735} {"adam_lr": 6.572141796596376e-07, "dt_ms": 155.138969, "global_tokens_per_sec": 105608.539628, "grad_norm": 35.25, "loss": 3.048886, "lr_stage": "cosine_decay", "mfu": 60.665616, "stage_name": "permutation_cycle", "status": "training", "step": 1907, "tokens_per_sec_per_gpu": 105608.539628, "vram_gb": 20.709735} {"adam_lr": 6.431867463506048e-07, "dt_ms": 154.57654, "global_tokens_per_sec": 105992.798136, "grad_norm": 22.625, "loss": 2.305517, "lr_stage": "cosine_decay", "mfu": 60.886348, "stage_name": "permutation_cycle", "status": "training", "step": 1908, "tokens_per_sec_per_gpu": 105992.798136, "vram_gb": 20.709735} {"adam_lr": 6.293096624544304e-07, "dt_ms": 153.612137, "global_tokens_per_sec": 106658.238971, "grad_norm": 30.75, "loss": 1.26667, "lr_stage": "cosine_decay", "mfu": 61.268603, "stage_name": "permutation_cycle", "status": "training", "step": 1909, "tokens_per_sec_per_gpu": 106658.238971, "vram_gb": 20.709735} {"adam_lr": 6.15582970243117e-07, "dt_ms": 155.019999, "global_tokens_per_sec": 105689.589428, "grad_norm": 31.125, "loss": 2.14941, "lr_stage": "cosine_decay", "mfu": 60.712174, "stage_name": "permutation_cycle", "status": "training", "step": 1910, "tokens_per_sec_per_gpu": 105689.589428, "vram_gb": 20.709735} {"adam_lr": 6.02006711530545e-07, "dt_ms": 154.811382, "global_tokens_per_sec": 105832.011557, "grad_norm": 23.5, "loss": 2.196718, "lr_stage": "cosine_decay", "mfu": 60.793986, "stage_name": "permutation_cycle", "status": "training", "step": 1911, "tokens_per_sec_per_gpu": 105832.011557, "vram_gb": 20.709735} {"adam_lr": 5.885809276723608e-07, "dt_ms": 154.798746, "global_tokens_per_sec": 105840.650598, "grad_norm": 30.375, "loss": 1.705239, "lr_stage": "cosine_decay", "mfu": 60.798949, "stage_name": "permutation_cycle", "status": "training", "step": 1912, "tokens_per_sec_per_gpu": 105840.650598, "vram_gb": 20.709735} {"adam_lr": 5.753056595658224e-07, "dt_ms": 153.130293, "global_tokens_per_sec": 106993.852689, "grad_norm": 30.0, "loss": 1.598152, "lr_stage": "cosine_decay", "mfu": 61.461393, "stage_name": "permutation_cycle", "status": "training", "step": 1913, "tokens_per_sec_per_gpu": 106993.852689, "vram_gb": 20.709735} {"adam_lr": 5.621809476497098e-07, "dt_ms": 155.681372, "global_tokens_per_sec": 105240.593799, "grad_norm": 28.75, "loss": 2.063905, "lr_stage": "cosine_decay", "mfu": 60.454253, "stage_name": "permutation_cycle", "status": "training", "step": 1914, "tokens_per_sec_per_gpu": 105240.593799, "vram_gb": 20.709735} {"adam_lr": 5.492068319041588e-07, "dt_ms": 155.30777, "global_tokens_per_sec": 105493.756196, "grad_norm": 20.5, "loss": 2.345274, "lr_stage": "cosine_decay", "mfu": 60.59968, "stage_name": "permutation_cycle", "status": "training", "step": 1915, "tokens_per_sec_per_gpu": 105493.756196, "vram_gb": 20.709735} {"adam_lr": 5.363833518505834e-07, "dt_ms": 153.790474, "global_tokens_per_sec": 106534.556923, "grad_norm": 25.5, "loss": 2.313034, "lr_stage": "cosine_decay", "mfu": 61.197556, "stage_name": "permutation_cycle", "status": "training", "step": 1916, "tokens_per_sec_per_gpu": 106534.556923, "vram_gb": 20.709735} {"adam_lr": 5.237105465515257e-07, "dt_ms": 153.455734, "global_tokens_per_sec": 106766.945398, "grad_norm": 34.5, "loss": 2.494361, "lr_stage": "cosine_decay", "mfu": 61.331048, "stage_name": "permutation_cycle", "status": "training", "step": 1917, "tokens_per_sec_per_gpu": 106766.945398, "vram_gb": 20.709735} {"adam_lr": 5.111884546105506e-07, "dt_ms": 153.75042, "global_tokens_per_sec": 106562.310795, "grad_norm": 33.0, "loss": 1.111201, "lr_stage": "cosine_decay", "mfu": 61.213498, "stage_name": "permutation_cycle", "status": "training", "step": 1918, "tokens_per_sec_per_gpu": 106562.310795, "vram_gb": 20.709735} {"adam_lr": 4.988171141721232e-07, "dt_ms": 155.503273, "global_tokens_per_sec": 105361.126379, "grad_norm": 33.75, "loss": 2.50355, "lr_stage": "cosine_decay", "mfu": 60.523492, "stage_name": "permutation_cycle", "status": "training", "step": 1919, "tokens_per_sec_per_gpu": 105361.126379, "vram_gb": 20.709735} {"adam_lr": 4.865965629214819e-07, "dt_ms": 154.985428, "global_tokens_per_sec": 105713.164306, "grad_norm": 25.0, "loss": 1.831157, "lr_stage": "cosine_decay", "mfu": 60.725716, "stage_name": "permutation_cycle", "status": "training", "step": 1920, "tokens_per_sec_per_gpu": 105713.164306, "vram_gb": 20.709735} {"adam_lr": 4.745268380845602e-07, "dt_ms": 155.120134, "global_tokens_per_sec": 105621.362876, "grad_norm": 34.0, "loss": 2.382858, "lr_stage": "cosine_decay", "mfu": 60.672982, "stage_name": "permutation_cycle", "status": "training", "step": 1921, "tokens_per_sec_per_gpu": 105621.362876, "vram_gb": 20.709735} {"adam_lr": 4.6260797642782014e-07, "dt_ms": 154.408216, "global_tokens_per_sec": 106108.343027, "grad_norm": 40.5, "loss": 2.545915, "lr_stage": "cosine_decay", "mfu": 60.952722, "stage_name": "permutation_cycle", "status": "training", "step": 1922, "tokens_per_sec_per_gpu": 106108.343027, "vram_gb": 20.709735} {"adam_lr": 4.508400142581859e-07, "dt_ms": 155.572653, "global_tokens_per_sec": 105314.139107, "grad_norm": 39.5, "loss": 3.05124, "lr_stage": "cosine_decay", "mfu": 60.496501, "stage_name": "permutation_cycle", "status": "training", "step": 1923, "tokens_per_sec_per_gpu": 105314.139107, "vram_gb": 20.709735} {"adam_lr": 4.392229874229159e-07, "dt_ms": 155.409813, "global_tokens_per_sec": 105424.48827, "grad_norm": 28.75, "loss": 2.288961, "lr_stage": "cosine_decay", "mfu": 60.559889, "stage_name": "permutation_cycle", "status": "training", "step": 1924, "tokens_per_sec_per_gpu": 105424.48827, "vram_gb": 20.709735} {"adam_lr": 4.277569313094809e-07, "dt_ms": 153.978586, "global_tokens_per_sec": 106404.405993, "grad_norm": 35.0, "loss": 2.253107, "lr_stage": "cosine_decay", "mfu": 61.122792, "stage_name": "permutation_cycle", "status": "training", "step": 1925, "tokens_per_sec_per_gpu": 106404.405993, "vram_gb": 20.709735} {"adam_lr": 4.1644188084548063e-07, "dt_ms": 156.498194, "global_tokens_per_sec": 104691.304151, "grad_norm": 34.75, "loss": 2.332436, "lr_stage": "cosine_decay", "mfu": 60.13872, "stage_name": "permutation_cycle", "status": "training", "step": 1926, "tokens_per_sec_per_gpu": 104691.304151, "vram_gb": 20.709735} {"adam_lr": 4.052778704985216e-07, "dt_ms": 155.822992, "global_tokens_per_sec": 105144.945271, "grad_norm": 19.0, "loss": 2.723715, "lr_stage": "cosine_decay", "mfu": 60.399309, "stage_name": "permutation_cycle", "status": "training", "step": 1927, "tokens_per_sec_per_gpu": 105144.945271, "vram_gb": 20.709735} {"adam_lr": 3.9426493427611177e-07, "dt_ms": 155.559778, "global_tokens_per_sec": 105322.855227, "grad_norm": 23.625, "loss": 2.432499, "lr_stage": "cosine_decay", "mfu": 60.501507, "stage_name": "permutation_cycle", "status": "training", "step": 1928, "tokens_per_sec_per_gpu": 105322.855227, "vram_gb": 20.709735} {"adam_lr": 3.8340310572556603e-07, "dt_ms": 154.360294, "global_tokens_per_sec": 106141.285036, "grad_norm": 17.0, "loss": 2.453841, "lr_stage": "cosine_decay", "mfu": 60.971645, "stage_name": "permutation_cycle", "status": "training", "step": 1929, "tokens_per_sec_per_gpu": 106141.285036, "vram_gb": 20.709735} {"adam_lr": 3.7269241793390085e-07, "dt_ms": 154.576063, "global_tokens_per_sec": 105993.125103, "grad_norm": 20.875, "loss": 2.045282, "lr_stage": "cosine_decay", "mfu": 60.886536, "stage_name": "permutation_cycle", "status": "training", "step": 1930, "tokens_per_sec_per_gpu": 105993.125103, "vram_gb": 20.709735} {"adam_lr": 3.621329035277232e-07, "dt_ms": 154.325247, "global_tokens_per_sec": 106165.389906, "grad_norm": 24.625, "loss": 2.263064, "lr_stage": "cosine_decay", "mfu": 60.985492, "stage_name": "permutation_cycle", "status": "training", "step": 1931, "tokens_per_sec_per_gpu": 106165.389906, "vram_gb": 20.709735} {"adam_lr": 3.517245946731529e-07, "dt_ms": 154.66857, "global_tokens_per_sec": 105929.731206, "grad_norm": 21.25, "loss": 1.414216, "lr_stage": "cosine_decay", "mfu": 60.85012, "stage_name": "permutation_cycle", "status": "training", "step": 1932, "tokens_per_sec_per_gpu": 105929.731206, "vram_gb": 20.709735} {"adam_lr": 3.414675230757003e-07, "dt_ms": 155.352831, "global_tokens_per_sec": 105463.157037, "grad_norm": 28.875, "loss": 2.565172, "lr_stage": "cosine_decay", "mfu": 60.582102, "stage_name": "permutation_cycle", "status": "training", "step": 1933, "tokens_per_sec_per_gpu": 105463.157037, "vram_gb": 20.709735} {"adam_lr": 3.3136171998017775e-07, "dt_ms": 153.689861, "global_tokens_per_sec": 106604.299475, "grad_norm": 23.25, "loss": 1.843937, "lr_stage": "cosine_decay", "mfu": 61.237618, "stage_name": "permutation_cycle", "status": "training", "step": 1934, "tokens_per_sec_per_gpu": 106604.299475, "vram_gb": 20.709735} {"adam_lr": 3.214072161706272e-07, "dt_ms": 155.120373, "global_tokens_per_sec": 105621.200537, "grad_norm": 32.75, "loss": 1.82173, "lr_stage": "cosine_decay", "mfu": 60.672888, "stage_name": "permutation_cycle", "status": "training", "step": 1935, "tokens_per_sec_per_gpu": 105621.200537, "vram_gb": 20.709735} {"adam_lr": 3.1160404197018154e-07, "dt_ms": 157.04298, "global_tokens_per_sec": 104328.127114, "grad_norm": 37.25, "loss": 2.247257, "lr_stage": "cosine_decay", "mfu": 59.930097, "stage_name": "permutation_cycle", "status": "training", "step": 1936, "tokens_per_sec_per_gpu": 104328.127114, "vram_gb": 20.709735} {"adam_lr": 3.019522272410202e-07, "dt_ms": 155.806065, "global_tokens_per_sec": 105156.368858, "grad_norm": 25.5, "loss": 1.890321, "lr_stage": "cosine_decay", "mfu": 60.405871, "stage_name": "permutation_cycle", "status": "training", "step": 1937, "tokens_per_sec_per_gpu": 105156.368858, "vram_gb": 20.709735} {"adam_lr": 2.924518013842303e-07, "dt_ms": 155.771494, "global_tokens_per_sec": 105179.706431, "grad_norm": 24.375, "loss": 3.100361, "lr_stage": "cosine_decay", "mfu": 60.419277, "stage_name": "permutation_cycle", "status": "training", "step": 1938, "tokens_per_sec_per_gpu": 105179.706431, "vram_gb": 20.709735} {"adam_lr": 2.8310279333976785e-07, "dt_ms": 156.113386, "global_tokens_per_sec": 104949.360229, "grad_norm": 26.25, "loss": 2.104275, "lr_stage": "cosine_decay", "mfu": 60.286958, "stage_name": "permutation_cycle", "status": "training", "step": 1939, "tokens_per_sec_per_gpu": 104949.360229, "vram_gb": 20.709735} {"adam_lr": 2.7390523158633554e-07, "dt_ms": 154.560566, "global_tokens_per_sec": 106003.752635, "grad_norm": 34.75, "loss": 2.438605, "lr_stage": "cosine_decay", "mfu": 60.892641, "stage_name": "permutation_cycle", "status": "training", "step": 1940, "tokens_per_sec_per_gpu": 106003.752635, "vram_gb": 20.709735} {"adam_lr": 2.6485914414129396e-07, "dt_ms": 155.561924, "global_tokens_per_sec": 105321.402441, "grad_norm": 34.25, "loss": 2.246314, "lr_stage": "cosine_decay", "mfu": 60.500673, "stage_name": "permutation_cycle", "status": "training", "step": 1941, "tokens_per_sec_per_gpu": 105321.402441, "vram_gb": 20.709735} {"adam_lr": 2.5596455856058963e-07, "dt_ms": 158.033371, "global_tokens_per_sec": 103674.305618, "grad_norm": 32.75, "loss": 2.308185, "lr_stage": "cosine_decay", "mfu": 59.554517, "stage_name": "permutation_cycle", "status": "training", "step": 1942, "tokens_per_sec_per_gpu": 103674.305618, "vram_gb": 20.709735} {"adam_lr": 2.472215019386881e-07, "dt_ms": 154.337645, "global_tokens_per_sec": 106156.861762, "grad_norm": 31.125, "loss": 1.940863, "lr_stage": "cosine_decay", "mfu": 60.980593, "stage_name": "permutation_cycle", "status": "training", "step": 1943, "tokens_per_sec_per_gpu": 106156.861762, "vram_gb": 20.709735} {"adam_lr": 2.386300009084408e-07, "dt_ms": 155.517101, "global_tokens_per_sec": 105351.757873, "grad_norm": 27.375, "loss": 2.303162, "lr_stage": "cosine_decay", "mfu": 60.51811, "stage_name": "permutation_cycle", "status": "training", "step": 1944, "tokens_per_sec_per_gpu": 105351.757873, "vram_gb": 20.709735} {"adam_lr": 2.3019008164105738e-07, "dt_ms": 155.576468, "global_tokens_per_sec": 105311.55683, "grad_norm": 33.0, "loss": 2.723304, "lr_stage": "cosine_decay", "mfu": 60.495017, "stage_name": "permutation_cycle", "status": "training", "step": 1945, "tokens_per_sec_per_gpu": 105311.55683, "vram_gb": 20.709735} {"adam_lr": 2.219017698460002e-07, "dt_ms": 154.756069, "global_tokens_per_sec": 105869.838168, "grad_norm": 28.625, "loss": 2.6077, "lr_stage": "cosine_decay", "mfu": 60.815716, "stage_name": "permutation_cycle", "status": "training", "step": 1946, "tokens_per_sec_per_gpu": 105869.838168, "vram_gb": 20.709735} {"adam_lr": 2.1376509077089545e-07, "dt_ms": 157.475948, "global_tokens_per_sec": 104041.284865, "grad_norm": 27.5, "loss": 2.678399, "lr_stage": "cosine_decay", "mfu": 59.765324, "stage_name": "permutation_cycle", "status": "training", "step": 1947, "tokens_per_sec_per_gpu": 104041.284865, "vram_gb": 20.709735} {"adam_lr": 2.057800692014833e-07, "dt_ms": 158.793688, "global_tokens_per_sec": 103177.90477, "grad_norm": 20.375, "loss": 2.152402, "lr_stage": "cosine_decay", "mfu": 59.269365, "stage_name": "permutation_cycle", "status": "training", "step": 1948, "tokens_per_sec_per_gpu": 103177.90477, "vram_gb": 20.709735} {"adam_lr": 1.9794672946152336e-07, "dt_ms": 153.012037, "global_tokens_per_sec": 107076.543072, "grad_norm": 41.0, "loss": 2.292546, "lr_stage": "cosine_decay", "mfu": 61.508893, "stage_name": "permutation_cycle", "status": "training", "step": 1949, "tokens_per_sec_per_gpu": 107076.543072, "vram_gb": 20.709735} {"adam_lr": 1.9026509541272275e-07, "dt_ms": 155.469418, "global_tokens_per_sec": 105384.070101, "grad_norm": 30.875, "loss": 3.148677, "lr_stage": "cosine_decay", "mfu": 60.536672, "stage_name": "permutation_cycle", "status": "training", "step": 1950, "tokens_per_sec_per_gpu": 105384.070101, "vram_gb": 20.709735} {"adam_lr": 1.8273519045468035e-07, "dt_ms": 155.308008, "global_tokens_per_sec": 105493.594249, "grad_norm": 15.9375, "loss": 2.038296, "lr_stage": "cosine_decay", "mfu": 60.599587, "stage_name": "permutation_cycle", "status": "training", "step": 1951, "tokens_per_sec_per_gpu": 105493.594249, "vram_gb": 20.709735} {"adam_lr": 1.753570375247815e-07, "dt_ms": 154.989958, "global_tokens_per_sec": 105710.074585, "grad_norm": 29.75, "loss": 2.230345, "lr_stage": "cosine_decay", "mfu": 60.723941, "stage_name": "permutation_cycle", "status": "training", "step": 1952, "tokens_per_sec_per_gpu": 105710.074585, "vram_gb": 20.709735} {"adam_lr": 1.6813065909817016e-07, "dt_ms": 156.025887, "global_tokens_per_sec": 105008.216033, "grad_norm": 20.625, "loss": 2.396495, "lr_stage": "cosine_decay", "mfu": 60.320767, "stage_name": "permutation_cycle", "status": "training", "step": 1953, "tokens_per_sec_per_gpu": 105008.216033, "vram_gb": 20.709735} {"adam_lr": 1.6105607718764347e-07, "dt_ms": 158.270836, "global_tokens_per_sec": 103518.755741, "grad_norm": 37.75, "loss": 2.827044, "lr_stage": "cosine_decay", "mfu": 59.465163, "stage_name": "permutation_cycle", "status": "training", "step": 1954, "tokens_per_sec_per_gpu": 103518.755741, "vram_gb": 20.709735} {"adam_lr": 1.5413331334360182e-07, "dt_ms": 158.223391, "global_tokens_per_sec": 103549.797157, "grad_norm": 33.25, "loss": 2.0128, "lr_stage": "cosine_decay", "mfu": 59.482995, "stage_name": "permutation_cycle", "status": "training", "step": 1955, "tokens_per_sec_per_gpu": 103549.797157, "vram_gb": 20.709735} {"adam_lr": 1.4736238865398765e-07, "dt_ms": 157.335997, "global_tokens_per_sec": 104133.830472, "grad_norm": 26.5, "loss": 1.438136, "lr_stage": "cosine_decay", "mfu": 59.818486, "stage_name": "permutation_cycle", "status": "training", "step": 1956, "tokens_per_sec_per_gpu": 104133.830472, "vram_gb": 20.709735} {"adam_lr": 1.407433237442135e-07, "dt_ms": 159.017086, "global_tokens_per_sec": 103032.953308, "grad_norm": 37.75, "loss": 2.259735, "lr_stage": "cosine_decay", "mfu": 59.1861, "stage_name": "permutation_cycle", "status": "training", "step": 1957, "tokens_per_sec_per_gpu": 103032.953308, "vram_gb": 20.709735} {"adam_lr": 1.342761387770952e-07, "dt_ms": 158.584595, "global_tokens_per_sec": 103313.944386, "grad_norm": 27.25, "loss": 1.309747, "lr_stage": "cosine_decay", "mfu": 59.347512, "stage_name": "permutation_cycle", "status": "training", "step": 1958, "tokens_per_sec_per_gpu": 103313.944386, "vram_gb": 20.709735} {"adam_lr": 1.2796085345280207e-07, "dt_ms": 157.815456, "global_tokens_per_sec": 103817.46107, "grad_norm": 21.75, "loss": 2.029673, "lr_stage": "cosine_decay", "mfu": 59.636751, "stage_name": "permutation_cycle", "status": "training", "step": 1959, "tokens_per_sec_per_gpu": 103817.46107, "vram_gb": 20.709735} {"adam_lr": 1.2179748700879012e-07, "dt_ms": 158.143282, "global_tokens_per_sec": 103602.251069, "grad_norm": 24.75, "loss": 2.006665, "lr_stage": "cosine_decay", "mfu": 59.513126, "stage_name": "permutation_cycle", "status": "training", "step": 1960, "tokens_per_sec_per_gpu": 103602.251069, "vram_gb": 20.709735} {"adam_lr": 1.1578605821973565e-07, "dt_ms": 159.157038, "global_tokens_per_sec": 102942.353245, "grad_norm": 17.625, "loss": 2.411808, "lr_stage": "cosine_decay", "mfu": 59.134055, "stage_name": "permutation_cycle", "status": "training", "step": 1961, "tokens_per_sec_per_gpu": 102942.353245, "vram_gb": 20.709735} {"adam_lr": 1.0992658539750178e-07, "dt_ms": 159.042597, "global_tokens_per_sec": 103016.426592, "grad_norm": 28.125, "loss": 1.900523, "lr_stage": "cosine_decay", "mfu": 59.176606, "stage_name": "permutation_cycle", "status": "training", "step": 1962, "tokens_per_sec_per_gpu": 103016.426592, "vram_gb": 20.709735} {"adam_lr": 1.0421908639104971e-07, "dt_ms": 155.092001, "global_tokens_per_sec": 105640.52239, "grad_norm": 34.5, "loss": 2.69962, "lr_stage": "cosine_decay", "mfu": 60.683988, "stage_name": "permutation_cycle", "status": "training", "step": 1963, "tokens_per_sec_per_gpu": 105640.52239, "vram_gb": 20.709735} {"adam_lr": 9.866357858642205e-08, "dt_ms": 154.96254, "global_tokens_per_sec": 105728.778288, "grad_norm": 26.0, "loss": 2.398237, "lr_stage": "cosine_decay", "mfu": 60.734685, "stage_name": "permutation_cycle", "status": "training", "step": 1964, "tokens_per_sec_per_gpu": 105728.778288, "vram_gb": 20.709735} {"adam_lr": 9.3260078906654e-08, "dt_ms": 153.529882, "global_tokens_per_sec": 106715.381661, "grad_norm": 34.25, "loss": 2.577923, "lr_stage": "cosine_decay", "mfu": 61.301428, "stage_name": "permutation_cycle", "status": "training", "step": 1965, "tokens_per_sec_per_gpu": 106715.381661, "vram_gb": 20.709735} {"adam_lr": 8.800860381173448e-08, "dt_ms": 155.254602, "global_tokens_per_sec": 105529.882807, "grad_norm": 20.75, "loss": 1.950526, "lr_stage": "cosine_decay", "mfu": 60.620432, "stage_name": "permutation_cycle", "status": "training", "step": 1966, "tokens_per_sec_per_gpu": 105529.882807, "vram_gb": 20.709735} {"adam_lr": 8.290916929858394e-08, "dt_ms": 158.433676, "global_tokens_per_sec": 103412.358015, "grad_norm": 29.25, "loss": 1.523559, "lr_stage": "cosine_decay", "mfu": 59.404044, "stage_name": "permutation_cycle", "status": "training", "step": 1967, "tokens_per_sec_per_gpu": 103412.358015, "vram_gb": 20.709735} {"adam_lr": 7.796179090094891e-08, "dt_ms": 155.065775, "global_tokens_per_sec": 105658.389214, "grad_norm": 37.75, "loss": 2.24515, "lr_stage": "cosine_decay", "mfu": 60.694251, "stage_name": "permutation_cycle", "status": "training", "step": 1968, "tokens_per_sec_per_gpu": 105658.389214, "vram_gb": 20.709735} {"adam_lr": 7.316648368941303e-08, "dt_ms": 155.07555, "global_tokens_per_sec": 105651.729055, "grad_norm": 25.75, "loss": 2.51767, "lr_stage": "cosine_decay", "mfu": 60.690425, "stage_name": "permutation_cycle", "status": "training", "step": 1969, "tokens_per_sec_per_gpu": 105651.729055, "vram_gb": 20.709735} {"adam_lr": 6.852326227130834e-08, "dt_ms": 153.357029, "global_tokens_per_sec": 106835.663882, "grad_norm": 15.5, "loss": 1.935857, "lr_stage": "cosine_decay", "mfu": 61.370523, "stage_name": "permutation_cycle", "status": "training", "step": 1970, "tokens_per_sec_per_gpu": 106835.663882, "vram_gb": 20.709735} {"adam_lr": 6.403214079069298e-08, "dt_ms": 154.652357, "global_tokens_per_sec": 105940.835995, "grad_norm": 28.375, "loss": 2.233205, "lr_stage": "cosine_decay", "mfu": 60.856499, "stage_name": "permutation_cycle", "status": "training", "step": 1971, "tokens_per_sec_per_gpu": 105940.835995, "vram_gb": 20.709735} {"adam_lr": 5.969313292830125e-08, "dt_ms": 155.265093, "global_tokens_per_sec": 105522.752728, "grad_norm": 33.5, "loss": 1.758141, "lr_stage": "cosine_decay", "mfu": 60.616336, "stage_name": "permutation_cycle", "status": "training", "step": 1972, "tokens_per_sec_per_gpu": 105522.752728, "vram_gb": 20.709735} {"adam_lr": 5.550625190150483e-08, "dt_ms": 156.442881, "global_tokens_per_sec": 104728.319588, "grad_norm": 23.5, "loss": 2.173707, "lr_stage": "cosine_decay", "mfu": 60.159983, "stage_name": "permutation_cycle", "status": "training", "step": 1973, "tokens_per_sec_per_gpu": 104728.319588, "vram_gb": 20.709735} {"adam_lr": 5.1471510464268236e-08, "dt_ms": 154.201269, "global_tokens_per_sec": 106250.74677, "grad_norm": 26.25, "loss": 2.057283, "lr_stage": "cosine_decay", "mfu": 61.034524, "stage_name": "permutation_cycle", "status": "training", "step": 1974, "tokens_per_sec_per_gpu": 106250.74677, "vram_gb": 20.709735} {"adam_lr": 4.7588920907110094e-08, "dt_ms": 157.958746, "global_tokens_per_sec": 103723.284841, "grad_norm": 25.625, "loss": 1.798465, "lr_stage": "cosine_decay", "mfu": 59.582653, "stage_name": "permutation_cycle", "status": "training", "step": 1975, "tokens_per_sec_per_gpu": 103723.284841, "vram_gb": 20.709735} {"adam_lr": 4.385849505708084e-08, "dt_ms": 155.580282, "global_tokens_per_sec": 105308.974679, "grad_norm": 45.0, "loss": 2.582276, "lr_stage": "cosine_decay", "mfu": 60.493534, "stage_name": "permutation_cycle", "status": "training", "step": 1976, "tokens_per_sec_per_gpu": 105308.974679, "vram_gb": 20.709735} {"adam_lr": 4.0280244277701716e-08, "dt_ms": 156.205654, "global_tokens_per_sec": 104887.368449, "grad_norm": 29.125, "loss": 2.439008, "lr_stage": "cosine_decay", "mfu": 60.251347, "stage_name": "permutation_cycle", "status": "training", "step": 1977, "tokens_per_sec_per_gpu": 104887.368449, "vram_gb": 20.709735} {"adam_lr": 3.685417946894254e-08, "dt_ms": 155.822277, "global_tokens_per_sec": 105145.427908, "grad_norm": 36.5, "loss": 2.351291, "lr_stage": "cosine_decay", "mfu": 60.399586, "stage_name": "permutation_cycle", "status": "training", "step": 1978, "tokens_per_sec_per_gpu": 105145.427908, "vram_gb": 20.709735} {"adam_lr": 3.35803110671884e-08, "dt_ms": 155.575991, "global_tokens_per_sec": 105311.879608, "grad_norm": 24.625, "loss": 2.004797, "lr_stage": "cosine_decay", "mfu": 60.495203, "stage_name": "permutation_cycle", "status": "training", "step": 1979, "tokens_per_sec_per_gpu": 105311.879608, "vram_gb": 20.709735} {"adam_lr": 3.04586490452119e-08, "dt_ms": 155.391455, "global_tokens_per_sec": 105436.943312, "grad_norm": 40.5, "loss": 2.393606, "lr_stage": "cosine_decay", "mfu": 60.567044, "stage_name": "permutation_cycle", "status": "training", "step": 1980, "tokens_per_sec_per_gpu": 105436.943312, "vram_gb": 20.709735} {"adam_lr": 2.7489202912145406e-08, "dt_ms": 156.322479, "global_tokens_per_sec": 104808.982552, "grad_norm": 36.25, "loss": 3.135633, "lr_stage": "cosine_decay", "mfu": 60.206319, "stage_name": "permutation_cycle", "status": "training", "step": 1981, "tokens_per_sec_per_gpu": 104808.982552, "vram_gb": 20.709735} {"adam_lr": 2.467198171342e-08, "dt_ms": 156.593561, "global_tokens_per_sec": 104627.545841, "grad_norm": 31.25, "loss": 2.38757, "lr_stage": "cosine_decay", "mfu": 60.102095, "stage_name": "permutation_cycle", "status": "training", "step": 1982, "tokens_per_sec_per_gpu": 104627.545841, "vram_gb": 20.709735} {"adam_lr": 2.2006994030798754e-08, "dt_ms": 154.201269, "global_tokens_per_sec": 106250.74677, "grad_norm": 27.5, "loss": 2.757909, "lr_stage": "cosine_decay", "mfu": 61.034524, "stage_name": "permutation_cycle", "status": "training", "step": 1983, "tokens_per_sec_per_gpu": 106250.74677, "vram_gb": 20.709735} {"adam_lr": 1.949424798228239e-08, "dt_ms": 155.836821, "global_tokens_per_sec": 105135.615169, "grad_norm": 25.625, "loss": 2.383961, "lr_stage": "cosine_decay", "mfu": 60.39395, "stage_name": "permutation_cycle", "status": "training", "step": 1984, "tokens_per_sec_per_gpu": 105135.615169, "vram_gb": 20.709735} {"adam_lr": 1.7133751222137007e-08, "dt_ms": 156.925917, "global_tokens_per_sec": 104405.953761, "grad_norm": 30.625, "loss": 2.221481, "lr_stage": "cosine_decay", "mfu": 59.974804, "stage_name": "permutation_cycle", "status": "training", "step": 1985, "tokens_per_sec_per_gpu": 104405.953761, "vram_gb": 20.709735} {"adam_lr": 1.4925510940844156e-08, "dt_ms": 154.59466, "global_tokens_per_sec": 105980.374876, "grad_norm": 27.875, "loss": 2.311448, "lr_stage": "cosine_decay", "mfu": 60.879212, "stage_name": "permutation_cycle", "status": "training", "step": 1986, "tokens_per_sec_per_gpu": 105980.374876, "vram_gb": 20.709735} {"adam_lr": 1.286953386508416e-08, "dt_ms": 156.772852, "global_tokens_per_sec": 104507.890217, "grad_norm": 29.0, "loss": 1.583355, "lr_stage": "cosine_decay", "mfu": 60.03336, "stage_name": "permutation_cycle", "status": "training", "step": 1987, "tokens_per_sec_per_gpu": 104507.890217, "vram_gb": 20.709735} {"adam_lr": 1.096582625772502e-08, "dt_ms": 155.480623, "global_tokens_per_sec": 105376.474946, "grad_norm": 18.0, "loss": 2.487581, "lr_stage": "cosine_decay", "mfu": 60.532309, "stage_name": "permutation_cycle", "status": "training", "step": 1988, "tokens_per_sec_per_gpu": 105376.474946, "vram_gb": 20.709735} {"adam_lr": 9.21439391778911e-09, "dt_ms": 155.016899, "global_tokens_per_sec": 105691.702609, "grad_norm": 22.0, "loss": 1.833334, "lr_stage": "cosine_decay", "mfu": 60.713388, "stage_name": "permutation_cycle", "status": "training", "step": 1989, "tokens_per_sec_per_gpu": 105691.702609, "vram_gb": 20.709735} {"adam_lr": 7.615242180436522e-09, "dt_ms": 155.671835, "global_tokens_per_sec": 105247.041032, "grad_norm": 31.5, "loss": 2.343849, "lr_stage": "cosine_decay", "mfu": 60.457957, "stage_name": "permutation_cycle", "status": "training", "step": 1990, "tokens_per_sec_per_gpu": 105247.041032, "vram_gb": 20.709735} {"adam_lr": 6.1683759169706146e-09, "dt_ms": 154.907465, "global_tokens_per_sec": 105766.368341, "grad_norm": 18.625, "loss": 2.160851, "lr_stage": "cosine_decay", "mfu": 60.756278, "stage_name": "permutation_cycle", "status": "training", "step": 1991, "tokens_per_sec_per_gpu": 105766.368341, "vram_gb": 20.709735} {"adam_lr": 4.873799534788059e-09, "dt_ms": 155.585289, "global_tokens_per_sec": 105305.585799, "grad_norm": 31.875, "loss": 2.769798, "lr_stage": "cosine_decay", "mfu": 60.491587, "stage_name": "permutation_cycle", "status": "training", "step": 1992, "tokens_per_sec_per_gpu": 105305.585799, "vram_gb": 20.709735} {"adam_lr": 3.731516977401039e-09, "dt_ms": 155.761003, "global_tokens_per_sec": 105186.790226, "grad_norm": 27.625, "loss": 2.553073, "lr_stage": "cosine_decay", "mfu": 60.423346, "stage_name": "permutation_cycle", "status": "training", "step": 1993, "tokens_per_sec_per_gpu": 105186.790226, "vram_gb": 20.709735} {"adam_lr": 2.741531724392843e-09, "dt_ms": 156.055689, "global_tokens_per_sec": 104988.162366, "grad_norm": 27.625, "loss": 2.435014, "lr_stage": "cosine_decay", "mfu": 60.309247, "stage_name": "permutation_cycle", "status": "training", "step": 1994, "tokens_per_sec_per_gpu": 104988.162366, "vram_gb": 20.709735} {"adam_lr": 1.903846791434516e-09, "dt_ms": 154.431581, "global_tokens_per_sec": 106092.289162, "grad_norm": 36.0, "loss": 2.312826, "lr_stage": "cosine_decay", "mfu": 60.9435, "stage_name": "permutation_cycle", "status": "training", "step": 1995, "tokens_per_sec_per_gpu": 106092.289162, "vram_gb": 20.709735} {"adam_lr": 1.2184647302626583e-09, "dt_ms": 155.262232, "global_tokens_per_sec": 105524.697199, "grad_norm": 28.375, "loss": 2.080512, "lr_stage": "cosine_decay", "mfu": 60.617453, "stage_name": "permutation_cycle", "status": "training", "step": 1996, "tokens_per_sec_per_gpu": 105524.697199, "vram_gb": 20.709735} {"adam_lr": 6.853876286627702e-10, "dt_ms": 155.617952, "global_tokens_per_sec": 105283.48274, "grad_norm": 30.125, "loss": 2.998689, "lr_stage": "cosine_decay", "mfu": 60.47889, "stage_name": "permutation_cycle", "status": "training", "step": 1997, "tokens_per_sec_per_gpu": 105283.48274, "vram_gb": 20.709735} {"adam_lr": 3.0461711048035415e-10, "dt_ms": 155.369282, "global_tokens_per_sec": 105451.990339, "grad_norm": 31.875, "loss": 2.754579, "lr_stage": "cosine_decay", "mfu": 60.575688, "stage_name": "permutation_cycle", "status": "training", "step": 1998, "tokens_per_sec_per_gpu": 105451.990339, "vram_gb": 20.709735} {"adam_lr": 7.615433561536378e-11, "dt_ms": 154.611588, "global_tokens_per_sec": 105968.771567, "grad_norm": 30.625, "loss": 2.064772, "lr_stage": "cosine_decay", "mfu": 60.872547, "stage_name": "permutation_cycle", "status": "training", "step": 1999, "tokens_per_sec_per_gpu": 105968.771567, "vram_gb": 20.709735} {"adam_lr": 0.0, "dt_ms": 154.653549, "global_tokens_per_sec": 105940.019388, "grad_norm": 22.625, "loss": 2.19963, "lr_stage": "cosine_decay", "mfu": 60.85603, "stage_name": "permutation_cycle", "status": "training", "step": 2000, "tokens_per_sec_per_gpu": 105940.019388, "vram_gb": 20.709735}