PyTorch native quantization and sparsity for training and inference - View it on GitHub