torchao

Also written: torch.ao.quantization

A library whose `quantize_` function applies int8 quantization to a model. It replaces PyTorch's deprecated `torch.ao.quantization`.