Sparsity-aware deep learning inference runtime for CPUs

Neural network model repository for highly sparse and sparse-quantized models with matching sparsification recipes

Flax is a neural network library for JAX that is designed for flexibility.

PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT

FFCV: Fast Forward Computer Vision (and other ML workloads!)

One Metrics Library to Rule Them All!

Libraries for applying sparsification recipes to neural networks with a few lines of code, enabling faster and smaller models