AITemplate is a Python framework which renders neural network into high performance CUDA/HIP C++ code. Specialized for FP16 TensorCore (NVIDIA GPU) and MatrixCore (AMD GPU) inference.

cuSignal - RAPIDS Signal Processing Library

An efficient C++17 GPU numerical computing library with Python-like syntax

HugeCTR is a high efficiency GPU framework designed for Click-Through-Rate (CTR) estimating training

Development repository for the Triton language and compiler

cudnn_frontend provides a c++ wrapper for the cudnn backend API and samples on how to use it

blingful character graphics/TUI library. definitely not curses.

CUDA Templates for Linear Algebra Subroutines

cuASR: CUDA Algebra for Semirings

Power meter for a bike that computes power, cadence, and resistance similar to a peloton bike.

Thin, unified, C++-flavored wrappers for the CUDA APIs

An Open Source Machine Learning Framework for Everyone

A PyTorch Extension: Tools for easy mixed precision and distributed training in Pytorch

The compiler is available for download. Get it!

ossia score, an interactive sequencer for the intermedia arts

Open-Source Vulkan C++ API

An open-source C++ library developed and used at Facebook.