lxq2t

atomic_queue

C++ lockless queue.

HVM

A massively parallel, optimal functional runtime in Rust

nvbench

CUDA Kernel Benchmarking Library

tensornetwork.org

Source for The Tensor Network open-source review article

v2rayNG

A V2Ray client for Android, support Xray core and v2fly core

antigen

The plugin manager for zsh.

polymer

Bridging polyhedral analysis tools to the MLIR framework

Polygeist

C/C++ frontend for MLIR. Also features polyhedral optimizations, parallel optimizations, and more!

FXdiv

C99/C++ header-only library for division via fixed-point multiplication by inverse

taskflow

A General-purpose Task-parallel Programming System using Modern C++

cutt

CUDA Tensor Transpose (cuTT) library

tpp-mlir

TPP experimentation on MLIR for linear algebra

ColossalAI

Making large AI models cheaper, faster and more accessible

mlir-tcp

Tensor Compute Primitives: Mid-level Intermediate Representation for Machine Learning Programs

kernl

Kernl lets you run PyTorch transformer models several times faster on GPU with a single line of code, and is designed to be easily hackable.

SHARK

SHARK - High Performance Machine Learning Distribution

mlir-hlo

iree-comparative-benchmark

MODel_opt

Memory Optimizations for Deep Learning (ICML 2023)

conan

Conan - The open-source C and C++ package manager

onnx-tensorflow

Tensorflow Backend for ONNX

tensorflow-onnx

Convert TensorFlow, Keras, Tensorflow.js and Tflite models to ONNX

CMSIS-NN

CMSIS-NN Library

stablehlo

Backward compatible ML compute opset inspired by HLO/MHLO

c-blosc

A blocking, shuffling and loss-less compression library that can be faster than `memcpy()`.

onnx

Open standard for machine learning interoperability

mindspore

MindSpore is a new open source deep learning training/inference framework that could be used for mobile, edge and cloud scenarios.

conan-center-index

Recipes for the ConanCenter repository

edotor.net

Your favourite Graphviz editor

TransformerEngine

A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs, to provide better performance with lower memory utilization in both training and inference.