Upload LoRA adapter (README written by author)

README.md

@@ -33,10 +33,10 @@ while intermediate reasoning (Chain-of-Thought) is masked.
 
 - Base model: Qwen/Qwen3-4B-Instruct-2507
 - Method: QLoRA (4-bit)
-- Max sequence length: 1024
-- Epochs: 5
-- Learning rate: 5e-05
-- LoRA: r=128, alpha=256
+- Max sequence length: 512
+- Epochs: 1
+- Learning rate: 1e-06
+- LoRA: r=64, alpha=128
 
 ## Usage
 

adapter_config.json

@@ -20,26 +20,26 @@
   "layers_pattern": null,
   "layers_to_transform": null,
   "loftq_config": {},
-  "lora_alpha": 256,
+  "lora_alpha": 128,
   "lora_bias": false,
-  "lora_dropout": 0.05,
+  "lora_dropout": 0.0,
   "megatron_config": null,
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
   "peft_version": "0.18.1",
   "qalora_group_size": 16,
-  "r": 128,
+  "r": 64,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "gate_proj",
     "o_proj",
-    "v_proj",
-    "down_proj",
+    "gate_proj",
     "k_proj",
-    "up_proj"
+    "up_proj",
+    "q_proj",
+    "v_proj",
+    "down_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3c90e53e8ea3eec4e3c1bff7deb869547c75f7c922ba2176724d7d3e10612363
-size 1057033224
+oid sha256:1429d34afe41ec1ba88fcda76c17e4199965d0b15a8d9e1c6cc38da2ef91aa9d
+size 528550256