123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113 |
- 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 any test models older than 1 week
- run: |
- python tests/scripts/remove_old_models.py --author bloom-testing --use_auth_token $BLOOM_TESTING_WRITE_TOKEN
- - name: Delete previous version of this model, if exists
- run: |
- export HF_TAG=$(python -c "import os; print(os.environ.get('GITHUB_HEAD_REF') or os.environ.get('GITHUB_REF_NAME'))")
- python -c "from huggingface_hub import delete_repo; delete_repo(token='$BLOOM_TESTING_WRITE_TOKEN', \
- name='test-bloomd-560m-$HF_TAG', organization='bloom-testing')" || true
- - name: Convert model and push to hub
- run: |
- export HF_TAG=$(python -c "import os; print(os.environ.get('GITHUB_HEAD_REF') or os.environ.get('GITHUB_REF_NAME'))")
- python -m cli.convert_model --model bigscience/bloom-560m --output_path ./converted_model \
- --output_repo bloom-testing/test-bloomd-560m-$HF_TAG --use_auth_token $BLOOM_TESTING_WRITE_TOKEN \
- --resize_token_embeddings 50000
- 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=$(python -c "import os; print(os.environ.get('GITHUB_HEAD_REF') or os.environ.get('GITHUB_REF_NAME'))")
- export MODEL_NAME=bloom-testing/test-bloomd-560m-$HF_TAG
- export REF_NAME=bigscience/bloom-560m
- python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 0:12 \
- --identity tests/test.id --host_maddrs /ip4/127.0.0.1/tcp/31337 --throughput 1 \
- --torch_dtype float32 --compression NONE --attn_cache_size 0.2GiB &> server1.log &
- SERVER1_PID=$!
-
- sleep 5 # wait for the first server to initialize DHT
-
- 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 12:22 \
- --initial_peers $INITIAL_PEERS --throughput 1 --torch_dtype float32 &> server2.log &
- SERVER2_PID=$!
- sleep 10 # wait for initial servers to declare blocks, then let server decide which blocks to serve
- python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 0:6 \
- --initial_peers $INITIAL_PEERS --throughput 1 --torch_dtype float32 &> server3.log &
- SERVER3_PID=$!
- python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --block_indices 4:16 \
- --torch_dtype float32 --initial_peers $INITIAL_PEERS --throughput 1 &> server4.log &
- SERVER4_PID=$!
-
- python -m cli.run_server --converted_model_name_or_path $MODEL_NAME --num_blocks 3 \
- --initial_peers $INITIAL_PEERS --throughput 1 --torch_dtype float32 &> server5.log &
- SERVER5_PID=$!
-
- tail -n 100 -f server*.log &
- LOGGER_PID=$!
- sleep 30 # wait for servers to download layers
-
- kill -0 $SERVER1_PID $SERVER2_PID $SERVER3_PID $SERVER4_PID $SERVER5_PID # ensure all servers survived init
-
- PYTHONPATH=. pytest tests --durations=0 --durations-min=1.0 -v
-
- kill -0 $SERVER1_PID $SERVER2_PID $SERVER3_PID $SERVER4_PID $SERVER5_PID # ensure all servers survived tests
-
- kill -s SIGINT $SERVER1_PID $SERVER2_PID $SERVER3_PID $SERVER4_PID $SERVER5_PID $LOGGER_PID
- echo "Done!"
|