Rust wrapper for Apple Matrix Coprocessor (AMX) instructions

Example models using DeepSpeed

Minimalist ML framework for Rust

DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective.

An open source framework that provides a simple, universal API for building distributed applications. Ray is packaged with RLlib, a scalable reinforcement learning library, and Tune, a scalable hyp…

Ongoing research training transformer language models at scale, including: BERT & GPT-2

🤗Transformers: State-of-the-art Natural Language Processing for Pytorch and TensorFlow 2.0.

Block-sparse primitives for PyTorch

Sparse multi-dimensional arrays for the PyData ecosystem

Exploring phase-space methods for collisionless dark matter simulations

For sorting CSV files on disk that do not fit into memory

SimTensor: Tensor data generator for evaluation of tensor factorization algorithms