Skip to content
forked from NVIDIA/apex

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

License

Notifications You must be signed in to change notification settings

Source-Roc/apex

 
 

Repository files navigation

Introduction

  • Ubuntu 22.04.5 LTS,Linux 5.15.0-157-generic
  • NVIDIA Driver Version: 570.195.03
  • Cuda compilation tools, release 12.8, V12.8.93 Build cuda_12.8.r12.8/compiler.35583870_0
  • CuDNN Version: 9.14.0

这个是我针对我的系统修改过的apex,修改如下:

  • 1、全面 删除了 VERSION_GE_1_1/1_3/1_5 定义
  • 2、修改 Tensor.type() => Tensor.options()
  • 3、针对有在 mlp.cpp 中有使用inputs.size() 作为for循环的判断条件时,同时使用 unsigned long 定义unsigned long i,删除了相关警告

原项目地址为:NVIDIA/apex

原项目README为:README

About

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

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 45.1%
  • Cuda 29.2%
  • C++ 24.6%
  • Other 1.1%