A CPU inference runtime for executing sparsified neural networks and language models with accelerator-like performance.

A repository and software interface for discovering and downloading pre-sparsified machine-learning models and optimization recipes.

An open-source library for pruning, quantizing, and otherwise compressing neural networks through reusable optimization recipes.

A graphical application for analyzing neural networks and applying performance-oriented sparsification workflows.

An open-source toolkit for applying quantization, sparsity, and compression algorithms to large language models.