Join the AMD Open Robotics Hackathon
• 16
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
Stabilizing Efficient Reasoning with Step-Level Advantage Selection
Dynamic Chunking Diffusion Transformer
Articles
Family of fully open 16B MoE LLM with 2.8B active params per token, trained on AMD Instinct™ MI300 & MI325 GPUs.
LLM-Compressor v0.10.0.2 quantized models for AMD EPYC CPU inference
TorchAO v0.17.0 quantized models for AMD EPYC CPU inference.
-
amd/Qwen2.5-VL-7B-Instruct-w4a16-tao-symgroup-torchao-v0.17.0
Image-Text-to-Text • Updated • 9 -
amd/Qwen3-VL-8B-Instruct-w4a16-tao-symgroup-torchao-v0.17.0
Image-Text-to-Text • 9B • Updated • 5 -
amd/Phi-4-reasoning-plus-w4a16-tao-symgroup-torchao-v0.17.0
Text Generation • 15B • Updated • 152 -
amd/Phi-4-reasoning-plus-da8w8-torchao-v0.17.0
Text Generation • 15B • Updated • 158
Liquid AI's Liquid Foundation (LFM2) ONNX based NPU models
Ryzen AI 1.7.1 models supporting context length up to 4K
-
amd/CodeLlama-7b-Instruct-hf_rai_1.7.1_npu_4K
Text Generation • Updated • 9 -
amd/DeepSeek-R1-Distill-Llama-8B_rai_1.7.1_npu_4K
Text Generation • Updated • 6 -
amd/DeepSeek-R1-Distill-Qwen-1.5B_rai_1.7.1_npu_4K
Text Generation • Updated • 9 -
amd/DeepSeek-R1-Distill-Qwen-7B_rai_1.7.1_npu_4K
Text Generation • Updated • 7
Action-controlled Interactive world model.
List will be updated
Hummingbird is a series of video generation models built on AMD Instinct™ GPUs, including text-to-video, image-to-videos models.
Model for Dell Pro AI studio
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 13 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-fp16-onnx-hybrid
Updated • 5 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5
Official Model Parameters for "Gumiho: A Hybrid Architecture to Prioritize Early Tokens in Speculative Decoding"
This collection contains AMD-Quark quantized OGA exported models for CPU execution
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 2 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 11 • 3 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Updated • 9 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 1
AMD-HybridLM is a family of post-trained, highly efficient hybrid models, designed to combine performance with speed and memory efficiency.
optimized image generation ONNX models for AMD Ryzen (TM) AI GPUs and Radeon Discrete GPUs
Models quantized by Quark and prepared for the OGA-based hybrid execution flow (Ryzen AI 1.3)
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 13 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-fp16-onnx-hybrid
Updated • 5 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5
AMD-OLMo are a series of 1 billion parameter language models trained by AMD on AMD Instinct™ MI250 GPUs based on OLMo.
Stable Diffusion (SD) and image generation models for AMD NPU
One Click Models List on HF
LLM-Compressor v0.11.0 quantized models for AMD EPYC CPU inference
TorchAO quantized models for AMD EPYC CPU inference. The inference stack includes vLLM (0.15.0 to 0.18.0), PyTorch 2.10, and zentorch 5.2.1.
-
amd/Llama-3.1-8B-Instruct-da8w8-torchao-v0.16.0
Text Generation • Updated • 15 • 1 -
amd/Qwen2.5-VL-7B-Instruct-da8w8-torchao-v0.16.0
Image-Text-to-Text • Updated • 9 -
amd/Qwen3-14B-Instruct-da8w8-torchao-v0.16.0
Text Generation • Updated • 14 -
amd/Phi-4-da8w8-torchao-v0.16.0
Text Generation • Updated • 107
Ryzen AI 1.7.1 models supporting context length up to 16K
-
amd/CodeLlama-7b-Instruct-hf_rai_1.7.1_npu_16K
Text Generation • Updated • 16 -
amd/DeepSeek-R1-Distill-Llama-8B_rai_1.7.1_npu_16K
Text Generation • Updated • 14 -
amd/DeepSeek-R1-Distill-Qwen-1.5B_rai_1.7.1_npu_16K
Text Generation • Updated • 8 -
amd/DeepSeek-R1-Distill-Qwen-7B_rai_1.7.1_npu_16K
Text Generation • Updated • 12
Ryzen AI 1.7.1 hybrid (NPU and GPU) execution models
Models are quantized using quark-0.9, transformers-4.50.0, OGA-0.7.1, ORT-1.21.1 followed by OGA-DML export.
PTPC model quantized by quark
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 2 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 11 • 3 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Updated • 9 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 1
Official Model Parameters for "PARD: Accelerating LLM Inference with Low-Cost PARallel Draft Model Adaptation"
Collection ONNX GenAI compatible Language Models to run on AMD Ryzen(TM) GPUs and Radeon Discrete GPUs
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 13 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-fp16-onnx-hybrid
Updated • 5 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5
Announcing Instella, a series of 3 billion parameter language models developed by AMD, trained from scratch on 128 Instinct MI300X GPUs.
Models quantized by Quark and prepared for the OGA-based NPU-only execution flow (Ryzen AI 1.3)
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 2 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 11 • 3 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Updated • 9 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 1
Nitro Diffusion is a series of efficient text-to-image diffusion models built on AMD Instinct™ GPUs.
ONNX Runtime generate() API based models quantized by Quark and
optimized for Ryzen AI Strix Point NPU
-
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 1 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Updated • 9 -
amd/Llama-2-7b-hf-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 103 -
amd/Llama-3-8B-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 11 • 2
Stable Diffusion (SD) and image generation models for AMD NPU
One Click Models List on HF
Family of fully open 16B MoE LLM with 2.8B active params per token, trained on AMD Instinct™ MI300 & MI325 GPUs.
LLM-Compressor v0.11.0 quantized models for AMD EPYC CPU inference
LLM-Compressor v0.10.0.2 quantized models for AMD EPYC CPU inference
TorchAO v0.17.0 quantized models for AMD EPYC CPU inference.
-
amd/Qwen2.5-VL-7B-Instruct-w4a16-tao-symgroup-torchao-v0.17.0
Image-Text-to-Text • Updated • 9 -
amd/Qwen3-VL-8B-Instruct-w4a16-tao-symgroup-torchao-v0.17.0
Image-Text-to-Text • 9B • Updated • 5 -
amd/Phi-4-reasoning-plus-w4a16-tao-symgroup-torchao-v0.17.0
Text Generation • 15B • Updated • 152 -
amd/Phi-4-reasoning-plus-da8w8-torchao-v0.17.0
Text Generation • 15B • Updated • 158
TorchAO quantized models for AMD EPYC CPU inference. The inference stack includes vLLM (0.15.0 to 0.18.0), PyTorch 2.10, and zentorch 5.2.1.
-
amd/Llama-3.1-8B-Instruct-da8w8-torchao-v0.16.0
Text Generation • Updated • 15 • 1 -
amd/Qwen2.5-VL-7B-Instruct-da8w8-torchao-v0.16.0
Image-Text-to-Text • Updated • 9 -
amd/Qwen3-14B-Instruct-da8w8-torchao-v0.16.0
Text Generation • Updated • 14 -
amd/Phi-4-da8w8-torchao-v0.16.0
Text Generation • Updated • 107
Liquid AI's Liquid Foundation (LFM2) ONNX based NPU models
Ryzen AI 1.7.1 models supporting context length up to 16K
-
amd/CodeLlama-7b-Instruct-hf_rai_1.7.1_npu_16K
Text Generation • Updated • 16 -
amd/DeepSeek-R1-Distill-Llama-8B_rai_1.7.1_npu_16K
Text Generation • Updated • 14 -
amd/DeepSeek-R1-Distill-Qwen-1.5B_rai_1.7.1_npu_16K
Text Generation • Updated • 8 -
amd/DeepSeek-R1-Distill-Qwen-7B_rai_1.7.1_npu_16K
Text Generation • Updated • 12
Ryzen AI 1.7.1 models supporting context length up to 4K
-
amd/CodeLlama-7b-Instruct-hf_rai_1.7.1_npu_4K
Text Generation • Updated • 9 -
amd/DeepSeek-R1-Distill-Llama-8B_rai_1.7.1_npu_4K
Text Generation • Updated • 6 -
amd/DeepSeek-R1-Distill-Qwen-1.5B_rai_1.7.1_npu_4K
Text Generation • Updated • 9 -
amd/DeepSeek-R1-Distill-Qwen-7B_rai_1.7.1_npu_4K
Text Generation • Updated • 7
Ryzen AI 1.7.1 hybrid (NPU and GPU) execution models
Action-controlled Interactive world model.
List will be updated
Hummingbird is a series of video generation models built on AMD Instinct™ GPUs, including text-to-video, image-to-videos models.
Models are quantized using quark-0.9, transformers-4.50.0, OGA-0.7.1, ORT-1.21.1 followed by OGA-DML export.
Model for Dell Pro AI studio
PTPC model quantized by quark
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 13 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-fp16-onnx-hybrid
Updated • 5 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 2 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 11 • 3 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Updated • 9 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 1
Official Model Parameters for "Gumiho: A Hybrid Architecture to Prioritize Early Tokens in Speculative Decoding"
Official Model Parameters for "PARD: Accelerating LLM Inference with Low-Cost PARallel Draft Model Adaptation"
This collection contains AMD-Quark quantized OGA exported models for CPU execution
Collection ONNX GenAI compatible Language Models to run on AMD Ryzen(TM) GPUs and Radeon Discrete GPUs
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 2 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 11 • 3 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Updated • 9 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 1
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 13 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-fp16-onnx-hybrid
Updated • 5 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5
Announcing Instella, a series of 3 billion parameter language models developed by AMD, trained from scratch on 128 Instinct MI300X GPUs.
AMD-HybridLM is a family of post-trained, highly efficient hybrid models, designed to combine performance with speed and memory efficiency.
optimized image generation ONNX models for AMD Ryzen (TM) AI GPUs and Radeon Discrete GPUs
Models quantized by Quark and prepared for the OGA-based NPU-only execution flow (Ryzen AI 1.3)
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 2 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 11 • 3 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Updated • 9 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 1
Models quantized by Quark and prepared for the OGA-based hybrid execution flow (Ryzen AI 1.3)
-
amd/Phi-3-mini-4k-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5 -
amd/Phi-3.5-mini-instruct-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 13 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-fp16-onnx-hybrid
Updated • 5 -
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-fp16-onnx-hybrid
Text Generation • Updated • 5
Nitro Diffusion is a series of efficient text-to-image diffusion models built on AMD Instinct™ GPUs.
AMD-OLMo are a series of 1 billion parameter language models trained by AMD on AMD Instinct™ MI250 GPUs based on OLMo.
ONNX Runtime generate() API based models quantized by Quark and
optimized for Ryzen AI Strix Point NPU
-
amd/Qwen1.5-7B-Chat-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 9 • 1 -
amd/Mistral-7B-Instruct-v0.3-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Updated • 9 -
amd/Llama-2-7b-hf-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 103 -
amd/Llama-3-8B-awq-g128-int4-asym-bf16-onnx-ryzen-strix
Text Generation • Updated • 11 • 2