Note: Pytorch Quantization development has transitioned to the TensorRT Model Optimizer. All developers are encouraged to use the TensorRT Model Optimizer to benefit from the latest advancements on quantization and compression. While the Pytorch Quantization code will remain available, it will no longer receive further development.
PyTorch-Quantization is a toolkit for training and evaluating PyTorch models with simulated quantization. Quantization can be added to the model automatically, or manually, allowing the model to be tuned for accuracy and performance. Quantization is compatible with NVIDIAs high performance integer kernels which leverage integer Tensor Cores. The quantized model can be exported to ONNX and imported by TensorRT 8.0 and later.
pip install pytorch-quantization --extra-index-url https://pypi.ngc.nvidia.comgit clone https://github.com/NVIDIA/TensorRT.git
cd tools/pytorch-quantizationInstall PyTorch and prerequisites
pip install -r requirements.txt
# for CUDA 10.2 users
pip install torch>=1.9.1
# for CUDA 11.1 users
pip install torch>=1.9.1+cu111Build and install pytorch-quantization
# Python version >= 3.7, GCC version >= 5.4 required
python setup.py installpytorch-quantization is preinstalled in NVIDIA NGC PyTorch container, e.g. nvcr.io/nvidia/pytorch:22.12-py3
- Pytorch Quantization Toolkit userguide
- Quantization Basics whitepaper