fabric

Author: Borda

.lightning/workflows/fabric.yml

@@ -0,0 +1,106 @@
+trigger:
+  push:
+    branches: ["master"]
+  pull_request:
+    branches: ["master"]
+
+timeout: "75" # minutes
+machine: "L4_X_2"
+parametrize:
+  matrix: {}
+  include:
+    # note that this is setting also all oldest requirements which is linked to Torch == 2.0
+    - image: "pytorchlightning/pytorch_lightning:base-cuda-py3.10-torch2.1-cuda12.1.1"
+      PACKAGE_NAME: "fabric"
+    - image: "pytorchlightning/pytorch_lightning:base-cuda-py3.12-torch2.7-cuda12.6.3"
+      PACKAGE_NAME: "fabric"
+    # - image: "pytorchlightning/pytorch_lightning:base-cuda-py3.12-torch2.7-cuda12.6.3"
+    #   PACKAGE_NAME: "fabric"
+    - image: "pytorchlightning/pytorch_lightning:base-cuda-py3.12-torch2.7-cuda12.6.3"
+      PACKAGE_NAME: "lightning"
+  exclude: []
+
+env:
+  FREEZE_REQUIREMENTS: "1"
+  RUN_ONLY_CUDA_TESTS: "1"
+
+run: |
+  whereis nvidia
+  nvidia-smi
+  python --version
+  pip --version
+  pip install -q fire wget packaging
+  set -ex
+
+  CUDA_VERSION="${image##*cuda}" # Remove everything up to and including "cuda"
+  echo "Using CUDA version: ${CUDA_VERSION}"
+  CUDA_VERSION_M_M="${cuda_version%.*}" # Get major.minor by removing the last dot and everything after
+  CUDA_VERSION_MM="${CUDA_VERSION_M_M//'.'/''}"
+  TORCH_URL="https://download.pytorch.org/whl/cu${CUDA_VERSION_MM}/torch_stable.html"
+  echo ${TORCH_URL}
+  COVERAGE_SOURCE=$(python -c 'n = "$(PACKAGE_NAME)" ; print(dict(fabric="lightning_fabric").get(n, n))')
+  echo "collecting coverage for: ${COVERAGE_SOURCE}"
+
+  if [ "${TORCH_VER}" == "2.1" ]; then
+    # Set oldest versions
+    python .actions/assistant.py replace_oldest_ver
+    pip install "cython<3.0" wheel  # for compatibility
+  fi
+
+  # Adjust torch versions in requirements files
+  PYTORCH_VERSION=$(python -c "import torch; print(torch.__version__.split('+')[0])")
+  pip install -q wget packaging
+  python -m wget https://raw.githubusercontent.com/Lightning-AI/utilities/main/scripts/adjust-torch-versions.py
+  for fpath in `ls requirements/**/*.txt`; do \
+    python ./adjust-torch-versions.py $fpath ${PYTORCH_VERSION}; \
+  done
+
+  if [ "${PACKAGE_NAME}" == "fabric" ]; then
+    pip install -U -q -r .actions/requirements.txt
+    python .actions/assistant.py copy_replace_imports --source_dir="./tests/tests_fabric" \
+      --source_import="lightning.fabric" \
+      --target_import="lightning_fabric"
+    python .actions/assistant.py copy_replace_imports --source_dir="./examples/fabric" \
+      --source_import="lightning.fabric" \
+      --target_import="lightning_fabric"
+  fi
+  
+  extra=$(python -c "print({'lightning': 'fabric-'}.get('$(PACKAGE_NAME)', ''))")
+  pip install -e ".[${extra}dev]" -U --upgrade-strategy=eager --extra-index-url="${TORCH_URL}"
+  
+  python requirements/collect_env_details.py
+  python -c "import torch ; mgpu = torch.cuda.device_count() ; assert mgpu >= 2, f'GPU: {mgpu}'"
+  python requirements/pytorch/check-avail-extras.py
+  python -c "import bitsandbytes"
+  
+  # Testing: Fabric doctests
+  if [ "${PACKAGE_NAME}" == "fabric" ]; then
+    cd src/
+    python -m pytest lightning_fabric
+    cd ..
+  fi
+  
+  cd tests/
+  # Testing: fabric standard
+  python -m coverage run --source ${COVERAGE_SOURCE} -m pytest tests_fabric/ -v --durations=50
+  
+  # Testing: fabric standalone
+  export PL_RUN_STANDALONE_TESTS=1
+  wget https://raw.githubusercontent.com/Lightning-AI/utilities/main/scripts/run_standalone_tests.sh
+  bash ./run_standalone_tests.sh "tests_fabric"
+  
+  python -m coverage report
+  python -m coverage xml
+  python -m coverage html
+
+  # TODO: enable coverage
+  #  # https://docs.codecov.com/docs/codecov-uploader
+  #  curl -Os https://uploader.codecov.io/latest/linux/codecov
+  #  chmod +x codecov
+  #  ./codecov --token=$(CODECOV_TOKEN) --commit=$(Build.SourceVersion) \
+  #    --flags=gpu,pytest,${COVERAGE_SOURCE} --name="GPU-coverage" --env=linux,azure
+  #  ls -l
+  
+  # Testing: fabric examples
+  bash run_fabric_examples.sh --accelerator=cuda --devices=1
+  bash run_fabric_examples.sh --accelerator=cuda --devices=2 --strategy ddp