facebookincubator/AITemplate
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.
PythonApache-2.0
Stargazers
- allensarkisyanLos Angeles, CA
- amit-cashify@reglobe
- antiBoson
- antinucleonFacebook
- brad-mengchiMeta
- brockoffdev@Giphy
- chaopliVMware
- chenyang78Meta
- Cobionix
- Craigacp@oracle Labs
- dhruv2601@exhuman-ai
- dzhulgakov@facebook
- gahdritzHarvard University
- ganlerUniversity of Illinois Urbana-Champaign
- hlu1@facebook
- junliumeAMD
- kerrmudgeonNVIDIA
- kflu
- kynk94NCSOFT Vision AI Lab
- mikeiovineNew York
- msaroufim@PyTorch
- p4perf4ceNowhere, Anywhere
- pbayliesDurham, NC
- piotrkawaPoland
- pkluska
- rainwangphy
- rishistypingParis, FR
- SelvamArulUniversität Bonn
- terrychenismU.S.
- thakkarVNVIDIA
- traversaroItalian Institute of Technology
- unverciftciMath & AI Institute
- vladbataev@yandex
- x86vk@NVIDIA
- yit-bUnited States
- zyan0Meta