12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061 |
- name: nv-accelerate-v100
- on:
- push:
- branches:
- - 'master'
- - 'staging**'
- paths-ignore:
- - 'docs/**'
- pull_request:
- paths-ignore:
- - 'docs/**'
- concurrency:
- group: ${{ github.workflow }}-${{ github.ref }}
- cancel-in-progress: true
- jobs:
- unit-tests:
- runs-on: [self-hosted, nvidia, cu111, v100]
- steps:
- - uses: actions/checkout@v2
- - name: environment
- run: |
- echo "JobID: $AISC_NODE_INSTANCE_ID"
- nvidia-smi
- which python
- python --version
- which nvcc
- nvcc --version
- pip install --upgrade pip
- pip uninstall --yes torch torchvision triton
- pip install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu111
- python -c "import torch; print('torch:', torch.__version__, torch)"
- python -c "import torch; print('CUDA available:', torch.cuda.is_available())"
- - name: Python environment
- run: |
- pip list
- - name: Install deepspeed
- run: |
- pip uninstall --yes deepspeed
- pip install .[dev,autotuning]
- ds_report
- - name: HF Accelerate tests
- run: |
- if [[ -d ./torch-extensions ]]; then rm -rf ./torch-extensions; fi
- git clone https://github.com/huggingface/accelerate
- cd accelerate
- # installing dependencies
- pip install .[testing]
- # force protobuf version due to issues
- pip install "protobuf<4.21.0"
- # tmp fix: force newer datasets version
- pip install "datasets>=2.0.0"
- pip list
- HF_DATASETS_CACHE=/blob/datasets_cache/ TRANSFORMERS_CACHE=/blob/transformers_cache/ TORCH_EXTENSIONS_DIR=./torch-extensions pytest --color=yes --durations=0 --verbose tests/deepspeed
|