diff --git a/.github/workflows/gpu-tests.yml b/.github/workflows/gpu-tests.yml index 81862e1f67b..0a72711fbdd 100644 --- a/.github/workflows/gpu-tests.yml +++ b/.github/workflows/gpu-tests.yml @@ -16,7 +16,7 @@ concurrency: group: gpu-tests-${{ github.ref_name }}-${{ !(github.ref_protected) || github.sha }} cancel-in-progress: true -# Cherry-picked from https://github.com/pytorch/test-infra/blob/main/.github/workflows/linux_job.yml +# Cherry-picked from https://github.com/pytorch/test-infra/blob/main/.github/workflows/linux_job_v2.yml jobs: gpu-tests: @@ -25,7 +25,7 @@ jobs: pytorch-channel: [pytorch, pytorch-nightly] fail-fast: false env: - DOCKER_IMAGE: "pytorch/conda-builder:cuda12.1" + DOCKER_IMAGE: "pytorch/almalinux-builder:cuda12.4" REPOSITORY: ${{ github.repository }} PR_NUMBER: ${{ github.event.pull_request.number }} runs-on: linux.8xlarge.nvidia.gpu @@ -40,7 +40,7 @@ jobs: echo "::endgroup::" - name: Checkout repository (pytorch/test-infra) - uses: actions/checkout@v3 + uses: actions/checkout@v4 with: # Support the use case where we need to checkout someone's fork repository: pytorch/test-infra @@ -55,7 +55,7 @@ jobs: docker-image: ${{ env.DOCKER_IMAGE }} - name: Checkout repository (${{ github.repository }}) - uses: actions/checkout@v3 + uses: actions/checkout@v4 with: # Support the use case where we need to checkout someone's fork repository: ${{ github.repository }} @@ -102,9 +102,9 @@ jobs: # Install PyTorch if [ "${{ matrix.pytorch-channel }}" == "pytorch" ]; then - pip install --upgrade torch torchvision --index-url https://download.pytorch.org/whl/cu121 + pip install --upgrade torch torchvision --index-url https://download.pytorch.org/whl/cu124 else - pip install --upgrade --pre torch torchvision --index-url https://download.pytorch.org/whl/nightly/cu121 + pip install --upgrade --pre torch torchvision --index-url https://download.pytorch.org/whl/nightly/cu124 fi python -c "import torch; print(torch.__version__, ', CUDA is available: ', torch.cuda.is_available()); exit(not torch.cuda.is_available())" @@ -139,7 +139,7 @@ jobs: - name: Run examples in container continue-on-error: false run: | - SCRIPT=$(cat << EOF + script=$(cat << EOF set -xe