Merge branch 'master' into feat/device_name

Author: GdoongMathew

.github/CONTRIBUTING.md

@@ -212,6 +212,8 @@ We welcome any useful contribution! For your convenience here's a recommended wo
    - [Test README](https://github.com/Lightning-AI/pytorch-lightning/blob/master/tests/README.md)
    - [CI/CD README](https://github.com/Lightning-AI/pytorch-lightning/tree/master/.github/workflows#readme)
 
+1. Once you have a PR opened (and thereby a PR number), please update the respective changelog for [fabric](https://github.com/Lightning-AI/pytorch-lightning/blob/master/src/lightning/fabric/CHANGELOG.md) or [pytorch](https://github.com/Lightning-AI/pytorch-lightning/blob/master/src/lightning/pytorch/CHANGELOG.md) subpackage depending on where you made your changes.
+
 1. When you feel ready for integrating your work, mark your PR "Ready for review".
 
    - Your code should be readable and follow the project's design principles.

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

@@ -40,23 +40,20 @@ jobs:
       matrix:
         os: [macOS-14, ubuntu-22.04, windows-2022]
         config:
-          # only run PyTorch latest
+          # Test unified "lightning" package with PyTorch 2.1-2.5
           - { pkg-name: "lightning", python-version: "3.10", pytorch-version: "2.1" }
           - { pkg-name: "lightning", python-version: "3.11", pytorch-version: "2.2.2" }
           - { pkg-name: "lightning", python-version: "3.11", pytorch-version: "2.3" }
           - { pkg-name: "lightning", python-version: "3.12.7", pytorch-version: "2.4.1" }
           - { pkg-name: "lightning", python-version: "3.12.7", pytorch-version: "2.5.1" }
 
-          # only run PyTorch latest with Python latest, use Fabric scope to limit dependency issues
+          # Test "fabric" package with PyTorch 2.6-2.9
           - { pkg-name: "fabric", python-version: "3.12.7", pytorch-version: "2.6" }
-
-          # "fabric" installs the standalone package
-          - { pkg-name: "fabric", python-version: "3.10", pytorch-version: "2.7" }
-
-          # adding recently cut Torch 2.7 - FUTURE
+          - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.7" }
           - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.8" }
+          - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.9" }
 
-          # "oldest" versions tests, only on minimum Python
+          # Test minimum supported versions (oldest)
           - { pkg-name: "fabric", pytorch-version: "2.1", requires: "oldest" }
     timeout-minutes: 25 # because of building grpcio on Mac
     env:

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

@@ -44,23 +44,20 @@ jobs:
       matrix:
         os: [macOS-14, ubuntu-22.04, windows-2022]
         config:
-          # only run PyTorch latest
+          # Test unified "lightning" package with PyTorch 2.1-2.5
           - { pkg-name: "lightning", python-version: "3.10", pytorch-version: "2.1" }
           - { pkg-name: "lightning", python-version: "3.11", pytorch-version: "2.2.2" }
           - { pkg-name: "lightning", python-version: "3.11", pytorch-version: "2.3" }
           - { pkg-name: "lightning", python-version: "3.12.7", pytorch-version: "2.4.1" }
           - { pkg-name: "lightning", python-version: "3.12.7", pytorch-version: "2.5.1" }
 
-          # only run PyTorch latest with Python latest, use PyTorch scope to limit dependency issues
+          # Test "pytorch" package with PyTorch 2.6-2.9
           - { pkg-name: "pytorch", python-version: "3.12.7", pytorch-version: "2.6" }
-
-          # "pytorch" installs the standalone package
-          - { pkg-name: "pytorch", python-version: "3.10", pytorch-version: "2.7" }
-
-          # adding recently cut Torch 2.7 - FUTURE
+          - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.7" }
           - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.8" }
+          - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.9" }
 
-          # "oldest" versions tests, only on minimum Python
+          # Test minimum supported versions (oldest)
           - { pkg-name: "pytorch", pytorch-version: "2.1", requires: "oldest" }
     timeout-minutes: 50
     env:

requirements/docs.txt

@@ -3,7 +3,7 @@ myst-parser >=0.18.1, <5.0.0
 nbsphinx >=0.8.5, <=0.9.7
 nbconvert >7.14, <7.17
 pandoc >=1.0, <=2.4
-docutils>=0.18.1,<=0.22.2
+docutils>=0.18.1,<=0.22.3
 sphinxcontrib-fulltoc >=1.0, <=1.2.0
 sphinxcontrib-mockautodoc
 sphinx-autobuild

requirements/pytorch/docs.txt

@@ -1,7 +1,7 @@
 -r ../docs.txt
 
 nbformat  # used for generate empty notebook
-ipython[notebook] <9.7.0
+ipython[notebook] <9.8.0
 setuptools<81.0  # workaround for `error in ipython setup command: use_2to3 is invalid.`
 
 onnxscript >= 0.2.2, < 0.6.0

requirements/pytorch/extra.txt

@@ -6,6 +6,6 @@ matplotlib>3.1, <3.11.0
 omegaconf >=2.2.3, <2.4.0
 hydra-core >=1.2.0, <1.4.0
 jsonargparse[signatures,jsonnet] >=4.39.0, <4.43.0
-rich >=12.3.0, <14.2.0
+rich >=12.3.0, <14.3.0
 tensorboardX >=2.2, <2.7.0  # min version is set by torch.onnx missing attribute
 bitsandbytes >=0.45.2,<0.47.0; platform_system != "Darwin"

src/lightning/fabric/CHANGELOG.md

@@ -25,7 +25,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 ### Fixed
 
--
+- Fixed issue in detecting MPIEnvironment with partial mpi4py installation ([#21353](https://github.com/Lightning-AI/pytorch-lightning/pull/21353))
+
+- Learning rate scheduler is stepped at the end of epoch when `on_train_batch_start` returns -1 ([#21296](https://github.com/Lightning-AI/pytorch-lightning/issues/21296)).
 
 
 ---

src/lightning/fabric/plugins/environments/mpi.py

@@ -73,7 +73,11 @@ def detect() -> bool:
         if not _MPI4PY_AVAILABLE:
             return False
 
-        from mpi4py import MPI
+        try:
+            # mpi4py may be installed without MPI being present
+            from mpi4py import MPI
+        except ImportError:
+            return False
 
         return MPI.COMM_WORLD.Get_size() > 1
 

src/lightning/pytorch/CHANGELOG.md

@@ -25,6 +25,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Added support for variable batch size in `ThroughputMonitor` ([#20236](https://github.com/Lightning-AI/pytorch-lightning/pull/20236))
 
 
+- Added `EMAWeightAveraging` callback that wraps Lightning's `WeightAveraging` class ([#21260](https://github.com/Lightning-AI/pytorch-lightning/pull/21260))
+
+
 - Added classmethod `Accelerator.device_name` so that users could see device information during setup. ([#21112](https://github.com/Lightning-AI/pytorch-lightning/pull/21112))
 
 

src/lightning/pytorch/callbacks/__init__.py

@@ -32,7 +32,7 @@
 from lightning.pytorch.callbacks.stochastic_weight_avg import StochasticWeightAveraging
 from lightning.pytorch.callbacks.throughput_monitor import ThroughputMonitor
 from lightning.pytorch.callbacks.timer import Timer
-from lightning.pytorch.callbacks.weight_averaging import WeightAveraging
+from lightning.pytorch.callbacks.weight_averaging import EMAWeightAveraging, WeightAveraging
 
 __all__ = [
     "BackboneFinetuning",
@@ -59,5 +59,6 @@
     "ThroughputMonitor",
     "Timer",
     "TQDMProgressBar",
+    "EMAWeightAveraging",
     "WeightAveraging",
 ]