Will Constable 346343e6b5 [DeviceMesh] Make _validate_tp_mesh_dim support 3D (#125763) ... Currently a 3D mesh with a submesh sliced out for TP is going to fail this check. According to @wanchaol in [this comment](https://github.com/pytorch/pytorch/pull/125250#discussion_r1586653669) it should be OK to remove these checks. Though I would appreciate a more careful review here, since I'm not too sure if there are other edge cases where these checks are important. Pull Request resolved: https://github.com/pytorch/pytorch/pull/125763 Approved by: https://github.com/wz337, https://github.com/wanchaol		2024-05-08 21:22:11 +00:00
..
__init__.py	[TP] Introduce Sequence Parallel Style for Laynorm/RMSNorm/Dropout (#121295)	2024-03-07 02:04:59 +00:00
_data_parallel_utils.py	[reland] pass shape/stride during tensor unflatten (#117340)	2024-01-13 19:33:47 +00:00
_utils.py	[DeviceMesh] Make _validate_tp_mesh_dim support 3D (#125763)	2024-05-08 21:22:11 +00:00
api.py	[TP] Add wildcard support (#122968)	2024-04-02 21:23:39 +00:00
ddp.py	[2D] Remove enable_2d_with_fsdp() API and make remove_enable_2d_with_fsdp private (#112473)	2023-11-16 01:14:00 +00:00
fsdp.py	[FSDP1][2D] Fix FSDP1 2D state_dict to use run_check=False (#123802)	2024-04-24 01:25:11 +00:00
input_reshard.py	fix: docstring error in torch/distributed module (#113241)	2023-11-09 19:10:20 +00:00
loss.py	[dtensor][TP] check funcol calls and improve doc for loss parallel (#121366)	2024-03-08 01:41:31 +00:00
style.py	[tp] add kwargs support to prepare_module_input (#124114)	2024-04-22 21:46:31 +00:00