Simon Fan ed83b0b70b [ddp] decouple python reducer from compilation mode (#147123) ... Current implementation reads as: we will only actually use the "python_reducer" config if the DDP forward is compiled. Otherwise, we will silently fallback to C++ reducer + no DDPOptimizer. I'm changing this behavior to always use the python reducer if the config is specified. Pull Request resolved: https://github.com/pytorch/pytorch/pull/147123 Approved by: https://github.com/fegin		2025-02-19 15:51:40 +00:00
..
__init__.py	remove allow-untyped-defs for torch/nn/parallel/__init__.py (#143437)	2024-12-18 08:50:37 +00:00
_functions.py	[5/N] Apply Ruff fixes and pyupgrade to Python 3.9 (#144205)	2025-01-15 04:00:47 +00:00
comm.py	[5/N] Apply Ruff fixes and pyupgrade to Python 3.9 (#144205)	2025-01-15 04:00:47 +00:00
data_parallel.py	[5/N] Apply Ruff fixes and pyupgrade to Python 3.9 (#144205)	2025-01-15 04:00:47 +00:00
distributed.py	[ddp] decouple python reducer from compilation mode (#147123)	2025-02-19 15:51:40 +00:00
parallel_apply.py	[5/N] Apply Ruff fixes and pyupgrade to Python 3.9 (#144205)	2025-01-15 04:00:47 +00:00
replicate.py	[5/N] Apply Ruff fixes and pyupgrade to Python 3.9 (#144205)	2025-01-15 04:00:47 +00:00
scatter_gather.py	[5/N] Apply Ruff fixes and pyupgrade to Python 3.9 (#144205)	2025-01-15 04:00:47 +00:00