sparsity

neural-compressor

+6

SOTA low-bit LLM quantization (INT8/FP8/MXFP8/INT4/MXFP4/NVFP4) & sparsity; leading model compression techniques on PyTorch, TensorFlow, and ONNX Runtime

pytorch/ao

+5

PyTorch native quantization for training and inference

nncf

+4

Neural Network Compression Framework for enhanced OpenVINO™ inference

model-optimization

+1

A toolkit to optimize ML models for deployment for Keras and TensorFlow, including quantization and pruning.