yaml snippet for building pytorch.

pytorch-on-circleci.yml

 build-pytorch:
    machine:
      image: windows-server-2019-nvidia:edge
    steps:
    - run:
        name: Download cudnn
        command: |
          $output = "cudnn.zip"
          $url = "https://storage.googleapis.com/circleci-image-file/cudnn-10.1-windows10-x64-v7.6.4.38.zip" #
          (New-Object System.Net.WebClient).DownloadFile($url, $output)
          Expand-Archive -Path $output -DestinationPath C:\cudnn
    - run:
        name: Copying cudnn libraries into CUDA's path
        command: |
          cp C:\cudnn\cuda\bin\cudnn64_7.dll 'C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin'
          cp C:\cudnn\cuda\include\cudnn.h 'C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\include'
          cp C:\cudnn\cuda\lib\x64\cudnn.lib 'C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\lib\x64'
    - run:
        name: Clone PyTorch
        command: |
          git clone --recursive https://github.com/pytorch/pytorch pytorch
          cd C:/Users/circleci/project/pytorch
          git submodule sync
          git submodule update --init --recursive
          cd ..
    - run:
        name: Build PyTorch
        shell: bash.exe
        command: |
          powershell.exe
          cd C:/Users/circleci/project/pytorch
          conda install numpy ninja pyyaml mkl mkl-include setuptools cmake cffi typing
          python setup.py install
          cd ..
    - run:
        name: Fetch a PyTorch example
        command: |
          git clone https://github.com/pytorch/examples
          cp -r examples/word_language_model/* ./
    - run:
        name: Build a machine learning model
        command: |
          python main.py --cuda --epochs 1
    resource_class: windows.gpu.small.prototype
    shell: powershell.exe -ExecutionPolicy Bypass