gudiandian / pytorch

Tensors and Dynamic neural networks in Python with strong GPU acceleration

This repository contains the source code implementation of Pytorch modified to profile runtime statistics for pytorch models.

Setup

You will need to compile the modified PyTorch code in this repositiry. Note that torchvision should be installed before compiling PyTorch if you want to train a model from torchvision:

pip install torchvision
git submodule update --init --recursive
python setup.py install

About

Tensors and Dynamic neural networks in Python with strong GPU acceleration

https://pytorch.org

Other

Languages

Language:C++ 52.6%Language:Python 32.1%Language:Cuda 6.9%Language:C 4.3%Language:CMake 1.6%Language:Shell 0.6%Language:Objective-C++ 0.6%Language:Assembly 0.4%Language:PureBasic 0.3%Language:Starlark 0.3%Language:Java 0.2%Language:Metal 0.1%Language:Dockerfile 0.1%Language:Batchfile 0.0%Language:Ruby 0.0%Language:HTML 0.0%Language:Yacc 0.0%Language:Objective-C 0.0%Language:PowerShell 0.0%Language:CSS 0.0%Language:LLVM 0.0%Language:Lua 0.0%Language:Makefile 0.0%Language:Vim Script 0.0%