Merge branch 'master' into feat/device_name

Author: GdoongMathew

.actions/assistant.py

@@ -459,6 +459,25 @@ def generate_docker_tags(
         tags = [f"{docker_project}:{tag}" for tag in tags]
         print(",".join(tags))
 
+    @staticmethod
+    def prune_pytest_as_errors(
+        pyproject_toml: str = "pyproject.toml", errors: tuple = ("FutureWarning", "DeprecationWarning")
+    ) -> None:
+        """Prune pytest warnings as errors from the pyproject.toml file."""
+        import tomlkit
+
+        with open(pyproject_toml, encoding="utf-8") as fopen:
+            content = fopen.read()
+        pyproject = tomlkit.parse(content)
+        filterwarnings = pyproject.get("tool", {}).get("pytest", {}).get("ini_options", {}).get("filterwarnings", [])
+        if not filterwarnings:
+            return
+        filterwarnings = [wrn for wrn in filterwarnings if not any(f"error::{err}" in wrn for err in errors)]
+        pyproject["tool"]["pytest"]["ini_options"]["filterwarnings"] = filterwarnings
+
+        with open(pyproject_toml, "w", encoding="utf-8") as fopen:
+            fopen.write(tomlkit.dumps(pyproject))
+
 
 if __name__ == "__main__":
     import jsonargparse

.azure/gpu-benchmarks.yml

@@ -100,7 +100,9 @@ jobs:
         workingDirectory: tests/
         displayName: "Testing: benchmarks"
 
-      - bash: bash run_standalone_tasks.sh
+      - bash: |
+          bash run_standalone_tasks.sh cpu
+          bash run_standalone_tasks.sh cuda
         workingDirectory: tests/parity_fabric
         # without succeeded this could run even if the job has already failed
         condition: and(succeeded(), eq(variables['PACKAGE_NAME'], 'fabric'))

.azure/gpu-tests-pytorch.yml

@@ -1,8 +1,3 @@
-# Python package
-# Create and test a Python package on multiple Python versions.
-# Add steps that analyze code, save the dist with the build record, publish to a PyPI-compatible index, and more:
-# https://docs.microsoft.com/azure/devops/pipelines/languages/python
-
 trigger:
   tags:
     include: ["*"]
@@ -24,18 +19,18 @@ pr:
       - "examples/run_pl_examples.sh"
       - "examples/pytorch/basics/backbone_image_classifier.py"
       - "examples/pytorch/basics/autoencoder.py"
+      - "requirements/fabric/**"
       - "requirements/pytorch/**"
       - "src/lightning/__init__.py"
       - "src/lightning/__setup__.py"
       - "src/lightning/__version__.py"
-      - "src/lightning/pytorch/**"
+      - "src/lightning_fabric/*"
+      - "src/lightning/fabric/**"
       - "src/pytorch_lightning/*"
+      - "src/lightning/pytorch/**"
       - "tests/tests_pytorch/**"
       - "tests/run_standalone_*.sh"
       - "pyproject.toml" # includes pytest config
-      - "requirements/fabric/**"
-      - "src/lightning/fabric/**"
-      - "src/lightning_fabric/*"
     exclude:
       - "requirements/*/docs.txt"
       - "*.md"

.github/checkgroup.yml

@@ -21,10 +21,10 @@ subprojects:
     checks:
       - "pl-cpu-guardian" # aggregated check for all cases
 
-  - id: "pytorch_lightning: Azure GPU"
+  - id: "pytorch_lightning: lit GPU"
     paths:
       - ".actions/*"
-      - ".azure/gpu-tests-pytorch.yml"
+      - ".lightning/workflows/pytorch.yml"
       # only the azure GPU workflow runs the examples
       # all examples don't need to be added because they aren't used in CI, but these are
       - "examples/run_pl_examples.sh"
@@ -47,13 +47,13 @@ subprojects:
       - "!*.md"
       - "!**/*.md"
     checks:
-      - "pytorch-lightning (GPUs) (testing Lightning | latest)"
-      - "pytorch-lightning (GPUs) (testing PyTorch | oldest)"
-      - "pytorch-lightning (GPUs) (testing PyTorch | latest)"
+      - "pytorch.yml / Lit Job (nvidia/cuda:12.1.1-runtime-ubuntu22.04, pytorch, 3.10, A100_X_2)"
+      - "pytorch.yml / Lit Job (nvidia/cuda:12.6.3-runtime-ubuntu22.04, lightning, 3.12, L4_X_2)"
+      - "pytorch.yml / Lit Job (nvidia/cuda:12.6.3-runtime-ubuntu22.04, pytorch, 3.12, L4_X_2)"
 
-  - id: "pytorch_lightning: Benchmarks"
+  - id: "Benchmarks"
     paths:
-      - ".azure/gpu-benchmarks.yml"
+      - ".lightning/workflows/benchmark.yml"
       - "requirements/fabric/**"
       - "requirements/pytorch/**"
       - "src/lightning/fabric/**"
@@ -65,7 +65,8 @@ subprojects:
       - "!*.md"
       - "!**/*.md"
     checks:
-      - "lightning.Benchmarks"
+      - "benchmark.yml / Lit Job (fabric)"
+      - "benchmark.yml / Lit Job (pytorch)"
 
   # Temporarily disabled
   #  - id: "pytorch-lightning: TPU workflow"
@@ -111,34 +112,6 @@ subprojects:
       - "docs-make (pytorch, doctest)"
       - "docs-make (pytorch, html)"
 
-  - id: "pytorch_lightning: Docker"
-    paths:
-      - ".actions/*"
-      - ".github/workflows/docker-build.yml"
-      - "dockers/**"
-      - "requirements/pytorch/**"
-      - "requirements/fabric/**"
-      - "setup.py"
-      - "!requirements/*/docs.txt"
-      - "!*.md"
-      - "!**/*.md"
-    checks:
-      - "build-cuda (3.10, 2.1.2, 12.1.1)"
-      - "build-cuda (3.11, 2.2.2, 12.1.1)"
-      - "build-cuda (3.11, 2.3.1, 12.1.1)"
-      - "build-cuda (3.11, 2.4.1, 12.1.1)"
-      - "build-cuda (3.12, 2.5.1, 12.1.1)"
-      - "build-cuda (3.12, 2.6.0, 12.4.1)"
-      #- "build-NGC"
-      - "build-pl (3.10, 2.1, 12.1.1)"
-      - "build-pl (3.11, 2.2, 12.1.1)"
-      - "build-pl (3.11, 2.3, 12.1.1)"
-      - "build-pl (3.11, 2.4, 12.1.1)"
-      - "build-pl (3.12, 2.5, 12.1.1)"
-      - "build-pl (3.12, 2.6, 12.4.1)"
-      - "build-pl (3.12, 2.7, 12.6.3)"
-      - "build-pl (3.12, 2.8, 12.6.3, true)"
-
   # SECTION: lightning_fabric
 
   - id: "lightning_fabric: CPU workflow"
@@ -156,10 +129,10 @@ subprojects:
     checks:
       - "fabric-cpu-guardian" # aggregated check for all cases
 
-  - id: "lightning_fabric: Azure GPU"
+  - id: "lightning_fabric: lit GPU"
     paths:
       - ".actions/*"
-      - ".azure/gpu-tests-fabric.yml"
+      - ".lightning/workflows/fabric.yml"
       - "examples/fabric/**"
       - "examples/run_fabric_examples.sh"
       - "requirements/fabric/**"
@@ -175,9 +148,9 @@ subprojects:
       - "!*.md"
       - "!**/*.md"
     checks:
-      - "lightning-fabric (GPUs) (testing Fabric | oldest)"
-      - "lightning-fabric (GPUs) (testing Fabric | latest)"
-      - "lightning-fabric (GPUs) (testing Lightning | latest)"
+      - "fabric.yml / Lit Job (nvidia/cuda:12.1.1-runtime-ubuntu22.04, fabric, 3.10, A100_X_2)"
+      - "fabric.yml / Lit Job (nvidia/cuda:12.6.3-runtime-ubuntu22.04, fabric, 3.12, L4_X_2)"
+      - "fabric.yml / Lit Job (nvidia/cuda:12.6.3-runtime-ubuntu22.04, lightning, 3.12, L4_X_2)"
 
   # Temporarily disabled
   #  - id: "lightning_fabric: TPU workflow"

.github/label-change.yml

@@ -54,12 +54,6 @@ package:
           - "src/lightning_fabric/*"
           - "src/pytorch_lightning/*"
 
-dockers:
-  - changed-files:
-      - any-glob-to-any-file:
-          - "dockers/**"
-          - ".github/workflows/docker-build.yml"
-
 dependencies:
   - changed-files:
       - any-glob-to-any-file:

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

@@ -99,11 +99,14 @@ jobs:
       - name: Set min. dependencies
         if: ${{ matrix.config.requires == 'oldest' }}
         run: |
-          cd requirements/fabric
           uv pip install -U "lightning-utilities[cli]"
-          python -m lightning_utilities.cli requirements set-oldest --req_files "['base.txt', 'strategies.txt', 'test.txt']"
+          python -m lightning_utilities.cli requirements set-oldest \
+            --req_files "['requirements/fabric/base.txt', 'requirements/fabric/strategies.txt', 'requirements/fabric/test.txt']"
           uv pip install "cython<3.0" wheel
           uv pip install "pyyaml==5.4" --no-build-isolation
+          # This script removes any line containing "error::FutureWarning" from pyproject.toml
+          uv pip install -r requirements/ci.txt
+          python .actions/assistant.py prune_pytest_as_errors
 
       - name: Adjust PyTorch versions in requirements files
         if: ${{ matrix.config.requires != 'oldest' }}

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

@@ -98,11 +98,14 @@ jobs:
       - name: Set min. dependencies
         if: ${{ matrix.config.requires == 'oldest' }}
         run: |
-          cd requirements/pytorch
           uv pip install -U "lightning-utilities[cli]"
-          python -m lightning_utilities.cli requirements set-oldest --req_files "['base.txt', 'extra.txt', 'strategies.txt', 'examples.txt', 'test.txt']"
+          python -m lightning_utilities.cli requirements set-oldest \
+            --req_files "['requirements/pytorch/base.txt', 'requirements/pytorch/extra.txt', 'requirements/pytorch/strategies.txt', 'requirements/pytorch/examples.txt', 'requirements/pytorch/test.txt']"
           uv pip install "cython<3.0" wheel
           uv pip install "pyyaml==5.4" --no-build-isolation
+          # This script removes any line containing "error::FutureWarning" from pyproject.toml
+          uv pip install -r requirements/ci.txt
+          python .actions/assistant.py prune_pytest_as_errors
 
       - name: Adjust PyTorch versions in requirements files
         if: ${{ matrix.config.requires != 'oldest' }}

.github/workflows/code-checks.yml

@@ -29,23 +29,32 @@ jobs:
     runs-on: ubuntu-22.04
     steps:
       - uses: actions/checkout@v5
-      - uses: actions/setup-python@v6
-        with:
-          python-version: "3.11"
 
-      - name: Mypy cache
-        uses: actions/cache@v4
+      - name: Install uv and set Python version
+        uses: astral-sh/setup-uv@v6
         with:
-          path: .mypy_cache
-          key: mypy-${{ hashFiles('requirements/typing.txt') }}
+          python-version: "3.11"
+          # TODO: Avoid activating environment like this
+          # see: https://github.com/astral-sh/setup-uv/tree/v6/?tab=readme-ov-file#activate-environment
+          activate-environment: true
+          enable-cache: true
 
       - name: Install dependencies
         env:
           FREEZE_REQUIREMENTS: 1
         timeout-minutes: 20
         run: |
-          pip install -e '.[pytorch-all,fabric-all]' -r requirements/typing.txt
-          pip list
+          uv pip install '.[pytorch-all,fabric-all]' -r requirements/typing.txt
+          uv pip list
+
+      - name: mypy cache
+        uses: actions/cache@v4
+        with:
+          path: .mypy_cache
+          key: mypy-${{ hashFiles('requirements/typing.txt') }}
 
       - name: Check typing
         run: mypy
+
+      - name: Minimize uv cache
+        run: uv cache prune --ci