Please note: This model is not the official implementation from AutoAWQ. Instead, it is a Qwen3 implementation featuring scales assigned per expert. To run this model, you will need to install and use the kishizaki-sci version of AutoAWQ in your environment.

git clone -b qwen3_moe https://github.com/kIshizaki-sci/AutoAWQ.git
pip install -e ./AutoAWQ

quantization : https://colab.research.google.com/drive/1FRWZBiWDYeNrVx6pOQEB24tJK8EGkZoR?usp=sharing
inference : https://colab.research.google.com/drive/1yN21HjT7v695voETVYa6qYsraeNW-8Cy?usp=sharing

Downloads last month
25
Safetensors
Model size
4.64B params
Tensor type
I32
·
F16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support