FB (Facebook) + GEMM (General Matrix-Matrix Multiplication)

Tensors and Dynamic neural networks in Python with strong GPU acceleration

PyTorch extensions for high performance and large scale training.

D2Go is a toolkit for efficient deep learning

Compiler for Neural Network hardware accelerators

Pytorch domain library for recommendation systems

A Python-level JIT compiler designed to make unmodified PyTorch programs faster.

Facebook AI Research Sequence-to-Sequence Toolkit written in Python.

CUDA Templates for Linear Algebra Subroutines

PArametrized Recommendation and Ai Model benchmark is a repository for development of numerous uBenchmarks as well as end to end nets for evaluation of training and inference platforms.

Tutel MoE: An Optimized Mixture-of-Experts Implementation

The Tensor Algebra SuperOptimizer for Deep Learning

Model summary in PyTorch similar to `model.summary()` in Keras

Collective communications library with various primitives for multi-machine training.

An implementation of a deep learning recommendation model (DLRM)

A set of examples around pytorch in Vision, Text, Reinforcement Learning, etc.

"Multi-Level Intermediate Representation" Compiler Infrastructure

Facebook AI Performance Evaluation Platform

Caffe2 is a lightweight, modular, and scalable deep learning framework.

Library targeting Intel Architecture (x86) for small, dense or sparse matrix multiplications, and small convolutions.

The Surprisingly ParalleL spArse Tensor Toolkit.

Official HPCG benchmark source code

GAP Benchmark Suite

Caffe for Sparse Convolutional Neural Networks

Julia package to read MatrixMarket file format

Implement Arnoldi and Lanczos methods for svds and eigs