petals/.github/workflows/run-tests.yaml

name: Tests

on:
  push:
    branches: [ main ]
  pull_request:

jobs:
  convert-model:
    runs-on: ubuntu-latest
    env:
      BLOOM_TESTING_WRITE_TOKEN: ${{ secrets.BLOOM_TESTING_WRITE_TOKEN }}
    timeout-minutes: 15
    steps:
      - uses: actions/checkout@v2
      - name: Set up Python
        uses: actions/setup-python@v2
        with:
          python-version: 3.9
      - name: Cache dependencies
        uses: actions/cache@v2
        with:
          path: ~/.cache/pip
          key: Key-v1-py3.9-${{ hashFiles('requirements.txt') }}-${{ hashFiles('requirements-dev.txt') }}
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -r requirements.txt
      - name: Delete previous model, if exists
        run: |
          export HF_TAG=$(python -c "import os; print(os.environ.get('GITHUB_BASE_REF') or os.environ.get('GITHUB_REF_NAME'))")
          python -c "from huggingface_hub import delete_repo; delete_repo(token='$BLOOM_TESTING_WRITE_TOKEN', \
          name='test-bloomd-350m-$HF_TAG', organization='bloom-testing')" || true
      - name: Convert model and push to hub
        run: |
          python -m cli.convert_model --model bigscience/bloom-350m  --output_path ./converted_model \
            --output_repo bloom-testing/test-bloomd-350m-$HF_TAG --use_auth_token $BLOOM_TESTING_WRITE_TOKEN


  run-tests:
    runs-on: ubuntu-latest
    needs: convert-model
    strategy:
      matrix:
        python-version: [ 3.7, 3.8, 3.9 ]
      fail-fast: false
    timeout-minutes: 15
    steps:
      - uses: actions/checkout@v2
      - name: Set up Python
        uses: actions/setup-python@v2
        with:
          python-version: ${{ matrix.python-version }}
      - name: Cache dependencies
        uses: actions/cache@v2
        with:
          path: ~/.cache/pip
          key: Key-v1-${{ matrix.python-version }}-${{ hashFiles('requirements.txt') }}-${{ hashFiles('requirements-dev.txt') }}
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -r requirements.txt
          pip install -r requirements-dev.txt
      - name: Test
        run: |
          export HF_TAG=$(python -c "import os; print(os.environ.get('GITHUB_BASE_REF') or os.environ.get('GITHUB_REF_NAME'))")
          export MODEL_NAME=bloom-testing/test-bloomd-350m-$HF_TAG
          python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 0:12 \
            --torch_dtype float32 --identity tests/test.id --host_maddrs /ip4/127.0.0.1/tcp/31337 --throughput 1 &
          SERVER1_PID=$!
          
          export INITIAL_PEERS=/ip4/127.0.0.1/tcp/31337/p2p/QmS9KwZptnVdB9FFV7uGgaTq4sEKBwcYeKZDfSpyKDUd1g
          # ^-- server 1 multiaddr is determined by --identity and --host_maddrs
          
          python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 12:24 \
            --torch_dtype float32 --initial_peers $INITIAL_PEERS --throughput 1 &> server2.log &
          SERVER2_PID=$!

          sleep 30  # wait for server to download layers
          
          # test individual blocks
          export PYTHONPATH=.
          BLOCK_UID=$MODEL_NAME.0 REF_NAME=$MODEL_NAME REF_INDEX=0 pytest tests/test_block_exact_match.py
          BLOCK_UID=$MODEL_NAME.19 REF_NAME=$MODEL_NAME REF_INDEX=19 pytest tests/test_block_exact_match.py

          REF_NAME=$MODEL_NAME pytest tests/test_chained_calls.py
          
          REF_NAME=bigscience/bloom-350m pytest tests/test_full_model.py
          
          kill -s SIGINT $SERVER1_PID $SERVER2_PID
          echo "Done!"