Merge branch 'master' into bugfix/strategy-type-validation

Author: awaelchli

.azure/gpu-tests-fabric.yml

@@ -45,15 +45,21 @@ jobs:
     pool: lit-rtx-3090
     variables:
       DEVICES: $( python -c 'print("$(Agent.Name)".split("_")[-1])' )
+      FREEZE_REQUIREMENTS: "1"
+      COVERAGE_SCOPE: $( python -c 'n = "$(PACKAGE_NAME)" ; print(dict(fabric="lightning_fabric").get(n, n))' )
     container:
       image: "pytorchlightning/pytorch_lightning:base-cuda-py3.9-torch1.13-cuda11.7.1"
       # default shm size is 64m. Increase it to avoid:
       # 'Error while creating shared memory: unhandled system error, NCCL version 2.7.8'
       options: "--gpus=all --shm-size=2gb"
-    # TODO: package parametrization
+    strategy:
+      matrix:
+        'pkg: Fabric':
+          PACKAGE_NAME: "fabric"
+        'pkg: Lightning':
+          PACKAGE_NAME: "lightning"
     workspace:
       clean: all
-
     steps:
     - bash: |
         echo "##vso[task.setvariable variable=CUDA_VISIBLE_DEVICES]$(DEVICES)"
@@ -62,6 +68,8 @@ jobs:
       displayName: 'set env. vars'
 
     - bash: |
+        echo $(DEVICES)
+        echo $(COVERAGE_SCOPE)
         echo $CUDA_VISIBLE_DEVICES
         echo $TORCH_URL
         lspci | egrep 'VGA|3D'
@@ -80,11 +88,7 @@ jobs:
         done
       displayName: 'Adjust dependencies'
 
-    - bash: |
-        pip install -e .[dev,strategies,examples] --find-links ${TORCH_URL}
-      env:
-        PACKAGE_NAME: "fabric"
-        FREEZE_REQUIREMENTS: "1"
+    - bash: pip install -e .[dev,strategies,examples] --find-links ${TORCH_URL}
       displayName: 'Install package & dependencies'
 
     - bash: |
@@ -94,17 +98,26 @@ jobs:
         python -c "import torch ; mgpu = torch.cuda.device_count() ; assert mgpu == 2, f'GPU: {mgpu}'"
       displayName: 'Env details'
 
+    - bash: python -m pytest lightning_fabric
+      workingDirectory: src
+      condition: eq(variables['PACKAGE_NAME'], 'fabric')
+      displayName: 'Testing: Fabric doctests'
+
     - bash: |
         pip install -q -r .actions/requirements.txt
         python .actions/assistant.py copy_replace_imports --source_dir="./tests" \
           --source_import="lightning.fabric,lightning.pytorch" \
           --target_import="lightning_fabric,pytorch_lightning"
-      displayName: 'Adjust tests'
+        python .actions/assistant.py copy_replace_imports --source_dir="./examples" \
+          --source_import="lightning.fabric,lightning.pytorch" \
+          --target_import="lightning_fabric,pytorch_lightning"
+      condition: eq(variables['PACKAGE_NAME'], 'fabric')
+      displayName: 'Adjust tests & examples'
 
-    - bash: python -m coverage run --source lightning_fabric -m pytest --ignore benchmarks -v --junitxml=$(Build.StagingDirectory)/test-results.xml --durations=50
+    - bash: python -m coverage run --source $(COVERAGE_SCOPE) -m pytest --ignore benchmarks -v --durations=50
+      workingDirectory: tests/tests_fabric
       env:
         PL_RUN_CUDA_TESTS: "1"
-      workingDirectory: tests/tests_fabric
       displayName: 'Testing: fabric standard'
       timeoutInMinutes: "10"
 
@@ -113,28 +126,21 @@ jobs:
       env:
         PL_RUN_CUDA_TESTS: "1"
         PL_STANDALONE_TESTS_SOURCE: "lightning_fabric"
+      condition: eq(variables['PACKAGE_NAME'], 'fabric')
       displayName: 'Testing: fabric standalone tests'
       timeoutInMinutes: "10"
 
     - bash: |
         python -m coverage report
         python -m coverage xml
         python -m coverage html
-        python -m codecov --token=$(CODECOV_TOKEN) --commit=$(Build.SourceVersion) --flags=gpu,pytest --name="GPU-coverage" --env=linux,azure
+        python -m codecov --token=$(CODECOV_TOKEN) --commit=$(Build.SourceVersion) \
+          --flags=gpu,pytest,$(COVERAGE_SCOPE) --name="GPU-coverage" --env=linux,azure
         ls -l
       workingDirectory: tests/tests_fabric
       displayName: 'Statistics'
 
-    - task: PublishTestResults@2
-      displayName: 'Publish test results'
-      inputs:
-        testResultsFiles: '$(Build.StagingDirectory)/test-results.xml'
-        testRunTitle: '$(Agent.OS) - $(Build.DefinitionName) - Python $(python.version)'
-      condition: succeededOrFailed()
-
     - script: |
-        # In order to run the examples, we need to substitute the meta package imports with the standalone package
-        python ../.actions/assistant.py copy_replace_imports --source_dir="./fabric" --source_import="lightning.fabric" --target_import="lightning_fabric.fabric"
         set -e
         bash run_fabric_examples.sh --accelerator=cuda --devices=1
         bash run_fabric_examples.sh --accelerator=cuda --devices=2 --strategy ddp

.azure/gpu-tests-pytorch.yml

@@ -40,29 +40,36 @@ pr:
 
 jobs:
   - job: testing
+    # how long to run the job before automatically cancelling
+    timeoutInMinutes: "80"
+    # how much time to give 'run always even if cancelled tasks' before stopping them
+    cancelTimeoutInMinutes: "2"
     strategy:
       matrix:
         'PyTorch & strategies':  # this uses torch 1.12 as not all strategies support 1.13 yet
           image: "pytorchlightning/pytorch_lightning:base-cuda-py3.9-torch1.12-cuda11.6.1"
           scope: "strategies"
-        'PyTorch - latest':
+          PACKAGE_NAME: "pytorch"
+        'PyTorch | latest':
           image: "pytorchlightning/pytorch_lightning:base-cuda-py3.9-torch1.13-cuda11.7.1"
           scope: ""
-    # how long to run the job before automatically cancelling
-    timeoutInMinutes: "80"
-    # how much time to give 'run always even if cancelled tasks' before stopping them
-    cancelTimeoutInMinutes: "2"
+          PACKAGE_NAME: "pytorch"
+        'Lightning pkg':
+          image: "pytorchlightning/pytorch_lightning:base-cuda-py3.9-torch1.13-cuda11.7.1"
+          scope: ""
+          PACKAGE_NAME: "lightning"
     pool: lit-rtx-3090
     variables:
       DEVICES: $( python -c 'print("$(Agent.Name)".split("_")[-1])' )
+      FREEZE_REQUIREMENTS: "1"
+      COVERAGE_SCOPE: $( python -c 'n = "$(PACKAGE_NAME)" ; print(dict(pytorch="pytorch_lightning").get(n, n))' )
     container:
       image: $(image)
       # default shm size is 64m. Increase it to avoid:
       # 'Error while creating shared memory: unhandled system error, NCCL version 2.7.8'
       options: "--gpus=all --shm-size=2gb"
     workspace:
       clean: all
-
     steps:
 
     - bash: |
@@ -75,6 +82,8 @@ jobs:
       displayName: 'set env. vars'
 
     - bash: |
+        echo $(DEVICES)
+        echo $(COVERAGE_SCOPE)
         echo $CUDA_VISIBLE_DEVICES
         echo $CUDA_VERSION_MM
         echo $PYTORCH_VERSION
@@ -95,9 +104,6 @@ jobs:
       displayName: 'Adjust dependencies'
 
     - bash: pip install -e .[extra,test,examples] --find-links ${TORCH_URL}
-      env:
-        PACKAGE_NAME: "pytorch"
-        FREEZE_REQUIREMENTS: "1"
       displayName: 'Install package & extras'
 
     - bash: pip uninstall -y -r requirements/pytorch/strategies.txt
@@ -132,19 +138,21 @@ jobs:
 
     - bash: python -m pytest pytorch_lightning
       workingDirectory: src
+      condition: eq(variables['PACKAGE_NAME'], 'pytorch')
       displayName: 'Testing: PyTorch doctests'
 
     - bash: |
         pip install -q -r .actions/requirements.txt
         python .actions/assistant.py copy_replace_imports --source_dir="./tests" \
           --source_import="lightning.fabric,lightning.pytorch" \
           --target_import="lightning_fabric,pytorch_lightning"
+      condition: eq(variables['PACKAGE_NAME'], 'pytorch')
       displayName: 'Adjust tests'
 
-    - bash: python -m coverage run --source pytorch_lightning -m pytest --ignore benchmarks -v --junitxml=$(Build.StagingDirectory)/test-results.xml --durations=50
+    - bash: python -m coverage run --source $(COVERAGE_SCOPE) -m pytest --ignore benchmarks -v --durations=50
+      workingDirectory: tests/tests_pytorch
       env:
         PL_RUN_CUDA_TESTS: "1"
-      workingDirectory: tests/tests_pytorch
       displayName: 'Testing: PyTorch standard'
       timeoutInMinutes: "35"
 
@@ -155,6 +163,7 @@ jobs:
         PL_RUN_CUDA_TESTS: "1"
         PL_STANDALONE_TESTS_SOURCE: "pytorch_lightning"
       displayName: 'Testing: PyTorch standalone tests'
+      condition: eq(variables['PACKAGE_NAME'], 'pytorch')
       timeoutInMinutes: "35"
 
     - bash: bash run_standalone_tasks.sh
@@ -169,18 +178,12 @@ jobs:
         python -m coverage report
         python -m coverage xml
         python -m coverage html
-        python -m codecov --token=$(CODECOV_TOKEN) --commit=$(Build.SourceVersion) --flags=gpu,pytest --name="GPU-coverage" --env=linux,azure
+        python -m codecov --token=$(CODECOV_TOKEN) --commit=$(Build.SourceVersion) \
+          --flags=gpu,pytest,$(COVERAGE_SCOPE) --name="GPU-coverage" --env=linux,azure
         ls -l
       workingDirectory: tests/tests_pytorch
       displayName: 'Statistics'
 
-    - task: PublishTestResults@2
-      displayName: 'Publish test results'
-      inputs:
-        testResultsFiles: '$(Build.StagingDirectory)/test-results.xml'
-        testRunTitle: '$(Agent.OS) - $(Build.DefinitionName) - Python $(python.version)'
-      condition: succeededOrFailed()
-
     - script: |
         set -e
         bash run_pl_examples.sh --trainer.accelerator=gpu --trainer.devices=1

.git-blame-ignore-revs

@@ -0,0 +1,16 @@
+# copyright Lightning AI team (#16647)
+770b7929255389503e907350e2380ff449229816
+# [App] Add Missing Copyright (#16625)
+2bab2bac01694680b6c3e4f3a19d5bcd361fcaf4
+# adding license (#16450)
+e4c3441b25a8c194a873c8850e9507771de7053c
+# update copyright in PL & Fabric (#16481)
+98f7696d1681974d34fad59c03b4b58d9524ed13
+# add copyr (#6661)
+d471fa30b3bf95cfe601014bac544754067241ca
+# add copyright to tests (#5143)
+35401706bf0b89b07bc1748fdc2df612baa2be2a
+# added copyright notices (#3062)
+f43028f3ae5333b4ef0b08cc34f5560736381962
+# copyright (#2710)
+44d85c12191098b9bad40536375b29b154d91a47

.github/checkgroup.yml

@@ -32,9 +32,6 @@ subprojects:
       - "pl-cpu (windows-2022, lightning, 3.10, 1.12)"
       - "pl-cpu (windows-2022, lightning, 3.10, 1.13)"
       - "pl-cpu (windows-2022, lightning, 3.8, 1.11, oldest)"
-      - "pl-cpu (slow, macOS-11, lightning, 3.8, 1.11)"
-      - "pl-cpu (slow, ubuntu-20.04, lightning, 3.8, 1.11)"
-      - "pl-cpu (slow, windows-2022, lightning, 3.8, 1.11)"
       - "pl-cpu (macOS-11, pytorch, 3.8, 1.13)"
       - "pl-cpu (ubuntu-20.04, pytorch, 3.8, 1.13)"
       - "pl-cpu (windows-2022, pytorch, 3.8, 1.13)"

.github/workflows/ci-examples-app.yml

@@ -75,10 +75,11 @@ jobs:
     - name: Install Lightning package & dependencies
       run: |
         # do not use `-e` because it will make both packages available since it adds `src` to `sys.path` automatically
-        pip install .[dev] -U -f ${TORCH_URL} -f ".wheels/"
+        pip install .[dev] -U -f ${TORCH_URL} -f ".wheels/" --prefer-binary
         pip list
     - name: Dump handy wheels
       if: github.event_name == 'push'
+      continue-on-error: true
       uses: ./.github/actions/pip-wheels
       with:
         torch-url: ${{ env.TORCH_URL }}

.github/workflows/ci-tests-app.yml

@@ -68,16 +68,18 @@ jobs:
         path: .wheels
         key: pip_wheels  # to not being cleand with crone
         enableCrossOsArchive: true
+    - run: ls -lh .wheels/
 
     - name: Switch PyTorch URL
       run: python -c "print('TORCH_URL=https://download.pytorch.org/whl/' + str('test/cpu/torch_test.html' if '${{matrix.release}}' == 'pre' else 'cpu/torch_stable.html'))" >> $GITHUB_ENV
 
     - name: Install package & dependencies
       run: |
-        pip install -e .[dev] -U -f ${TORCH_URL} -f ".wheels/"
+        pip install -e .[dev] -U -f ${TORCH_URL} -f ".wheels/" --prefer-binary
         pip list
     - name: Dump handy wheels
       if: github.event_name == 'push'
+      continue-on-error: true
       uses: ./.github/actions/pip-wheels
       with:
         torch-url: ${{ env.TORCH_URL }}

.github/workflows/ci-tests-fabric.yml

@@ -92,17 +92,19 @@ jobs:
         path: .wheels
         key: pip_wheels  # to not being cleand with crone
         enableCrossOsArchive: true
+    - run: ls -lh .wheels/
 
     - name: Switch PyTorch URL
       run: python -c "print('TORCH_URL=https://download.pytorch.org/whl/' + str('test/cpu/torch_test.html' if '${{matrix.release}}' == 'pre' else 'cpu/torch_stable.html'))" >> $GITHUB_ENV
 
     - name: Install package & dependencies
       run: |
-        pip install -e .[test] "pytest-timeout" -U -f ${TORCH_URL} -f ".wheels/"
-        pip install -r requirements/fabric/strategies.txt -f ".wheels/"
+        pip install -e .[test] "pytest-timeout" -U -f ${TORCH_URL} -f ".wheels/" --prefer-binary
+        pip install -r requirements/fabric/strategies.txt -f ".wheels/" --prefer-binary
         pip list
     - name: Dump handy wheels
       if: github.event_name == 'push'
+      continue-on-error: true
       uses: ./.github/actions/pip-wheels
       with:
         torch-url: ${{ env.TORCH_URL }}

.github/workflows/ci-tests-pytorch.yml

@@ -57,10 +57,6 @@ jobs:
           - {os: "macOS-11", pkg-name: "lightning", python-version: "3.8", pytorch-version: "1.11", requires: "oldest"}
           - {os: "ubuntu-20.04", pkg-name: "lightning", python-version: "3.8", pytorch-version: "1.11", requires: "oldest"}
           - {os: "windows-2022", pkg-name: "lightning", python-version: "3.8", pytorch-version: "1.11", requires: "oldest"}
-          # run test under SLOW label
-          - {type: "slow", os: "macOS-11", pkg-name: "lightning", python-version: "3.8", pytorch-version: "1.11"}
-          - {type: "slow", os: "ubuntu-20.04", pkg-name: "lightning", python-version: "3.8", pytorch-version: "1.11"}
-          - {type: "slow", os: "windows-2022", pkg-name: "lightning", python-version: "3.8", pytorch-version: "1.11"}
           # "pytorch" installs the standalone package
           - {os: "macOS-11", pkg-name: "pytorch", python-version: "3.8", pytorch-version: "1.13"}
           - {os: "ubuntu-20.04", pkg-name: "pytorch", python-version: "3.8", pytorch-version: "1.13"}
@@ -88,7 +84,6 @@ jobs:
         python .actions/assistant.py replace_oldest_ver
 
     - name: Pull legacy checkpoints
-      if: ${{ matrix.type != 'slow' }}
       run: bash .actions/pull_legacy_checkpoints.sh
 
     - name: Adjust PyTorch versions in requirements files
@@ -105,24 +100,22 @@ jobs:
         path: .wheels
         key: pip_wheels  # to not being cleand with crone
         enableCrossOsArchive: true
+    - run: ls -lh .wheels/
 
     - name: Env. variables
       run: |
         # Switch PyTorch URL
         python -c "print('TORCH_URL=https://download.pytorch.org/whl/' + str('test/cpu/torch_test.html' if '${{matrix.release}}' == 'pre' else 'cpu/torch_stable.html'))" >> $GITHUB_ENV
         # Switch coverage scope
         python -c "print('COVERAGE_SCOPE=' + str('lightning' if '${{matrix.pkg-name}}' == 'lightning' else 'pytorch_lightning'))" >> $GITHUB_ENV
-        # Whether to run slow tests
-        python -c "print('PL_RUN_SLOW_TESTS=' + str(int('${{ matrix.type }}' == 'slow')))" >> $GITHUB_ENV
-        # Decide timeout
-        python -c "print('TEST_TIMEOUT=' + str(120 if '${{ matrix.type }}' == 'slow' else 60))" >> $GITHUB_ENV
 
     - name: Install package & dependencies
       run: |
-        pip install -e .[extra,test] "pytest-timeout" --upgrade -f ${TORCH_URL} -f ".wheels/"
+        pip install -e .[extra,test] "pytest-timeout" -U -f ${TORCH_URL} -f ".wheels/" --prefer-binary
         pip list
     - name: Dump handy wheels
       if: github.event_name == 'push'
+      continue-on-error: true
       uses: ./.github/actions/pip-wheels
       with:
         torch-url: ${{ env.TORCH_URL }}
@@ -155,9 +148,7 @@ jobs:
       # NOTE: do not include coverage report here, see: https://github.com/nedbat/coveragepy/issues/1003
       run: |
         python -m coverage run --source ${{ env.COVERAGE_SCOPE }} \
-          -m pytest . -v \
-          --timeout=${TEST_TIMEOUT} --durations=50 \
-          --reruns 3 --reruns-delay 1
+          -m pytest . -v --timeout=60 --durations=50 --reruns 3 --reruns-delay 1
 
     - name: Statistics
       if: success()

docs/source-pytorch/api_references.rst

@@ -254,7 +254,6 @@ utilities
     data
     deepspeed
     distributed
-    finite_checks
     memory
     model_summary
     parsing

docs/source-pytorch/common/progress_bar.rst

@@ -22,7 +22,7 @@ The :class:`~pytorch_lightning.callbacks.TQDMProgressBar` uses the `tqdm <https:
 It prints to ``stdout`` and shows up to four different bars:
 
 - **sanity check progress:** the progress during the sanity check run
-- **main progress:** shows training + validation progress combined. It also accounts for multiple validation runs during training when :paramref:`~pytorch_lightning.trainer.trainer.Trainer.val_check_interval` is used.
+- **train progress:** shows the training progress. It will pause if validation starts and will resume when it ends, and also accounts for multiple validation runs during training when :paramref:`~pytorch_lightning.trainer.trainer.Trainer.val_check_interval` is used.
 - **validation progress:** only visible during validation; shows total progress over all validation datasets.
 - **test progress:** only active when testing; shows total progress over all test datasets.