Pytorch Quantization (torq
) is a Python package that provides quantization features:
- Post Training Quantization (PTQ)
- Quantization Aware Training (QAT)
You can easily extend custom features such as Quantized Module, Quantizer, Observer and Data Type for your work when needed.
We provide the Pytorch Quantization Python package for an easy installation with pip
:
pip install git+https://github.com/insuofficial/pytorch-quantization.git
Or you can just clone the repository and modify the main code for your simulation:
git clone https://github.com/insuofficial/pytorch-quantization.git
The code here is licensed Apache 2.0, as found in the LICENSE file.