pytorch/torch/distributed
Yanli Zhao 956bd03808 add ignored_states to FSDP/fully_shard (#102056)
Add 'ignored_states' that accepts either a list of ignored_parameters or a list of nn modules for FSDP model wrapper and fully_shard composable APIs, it is recommended to use 'ignored_states' over 'ignored_modules' moving forward

Pull Request resolved: https://github.com/pytorch/pytorch/pull/102056
Approved by: https://github.com/awgu
2023-05-24 18:36:48 +00:00
..
_composable add ignored_states to FSDP/fully_shard (#102056) 2023-05-24 18:36:48 +00:00
_shard Fix type annotation of torch.split (#100655) 2023-05-16 21:35:41 +00:00
_sharded_tensor
_sharding_spec Add __all__ for a few distributed modules plus a little typing (reland) (#84872) 2022-09-13 21:57:49 +00:00
_spmd [spmd] Enable data parallel to work with non 0 batch dim (#100073) 2023-05-24 17:55:10 +00:00
_tensor [dtensor] hide mesh validation check under init_process_group flag (#101996) 2023-05-23 18:17:54 +00:00
_tools Fix typos under torch/distributed directory (#95638) 2023-03-27 21:13:44 +00:00
algorithms Properly propagates checkpoint wrapper args and kwargs (#99791) 2023-05-03 23:19:21 +00:00
autograd [BE]: pyupgrade Python to 3.8 - imports and object inheritance only (#94308) 2023-02-07 21:10:56 +00:00
benchmarks [BE] [2/3] Rewrite super() calls in functorch and torch (#94588) 2023-02-10 21:16:33 +00:00
checkpoint [BE] Fix flake8 B027 errors - missing abstractmethod decorator (#100715) 2023-05-09 17:28:48 +00:00
elastic [BE] Fix flake8 B027 errors - missing abstractmethod decorator (#100715) 2023-05-09 17:28:48 +00:00
examples Fix typos under torch/distributed directory (#95638) 2023-03-27 21:13:44 +00:00
fsdp add ignored_states to FSDP/fully_shard (#102056) 2023-05-24 18:36:48 +00:00
launcher Convert logging f-strings to use % format, part four (#98705) 2023-04-11 13:17:59 +00:00
nn [PTD] Make input contiguous for _ReduceScatter (#101373) 2023-05-15 22:08:21 +00:00
optim [BE]: Bugfix functorch and some generic typing improvements (#101337) 2023-05-14 14:20:56 +00:00
pipeline Enable ruff in lintrunner (#99785) 2023-04-24 16:18:44 +00:00
rpc [BE]: enable PLE error codes in ruff and fix bugs (#101079) 2023-05-11 23:57:25 +00:00
tensor [dtensor] tensor ops to use strategy based sharding prop (#100607) 2023-05-11 02:47:20 +00:00
__init__.py [c10d] Record time spent for init_process_group, new_group, _store_based_barrier (#101912) 2023-05-24 09:36:34 +00:00
_composable_state.py [Composable API]Common _State parent class for composable and wrapper FSDP (#89147) 2022-12-13 23:58:01 +00:00
_functional_collectives.py Add top level function to check if running with deploy (#101420) 2023-05-16 16:05:49 +00:00
argparse_util.py
c10d_logger.py [c10d] Record time spent for init_process_group, new_group, _store_based_barrier (#101912) 2023-05-24 09:36:34 +00:00
constants.py
CONTRIBUTING.md
distributed_c10d.py [c10d] Record time spent for init_process_group, new_group, _store_based_barrier (#101912) 2023-05-24 09:36:34 +00:00
launch.py Fix typos under torch/distributed directory (#95638) 2023-03-27 21:13:44 +00:00
logging_handlers.py [C10D][BE] Add exception handlers to c10d collectives function (#87643) (#87988) 2022-10-29 04:38:34 +00:00
remote_device.py [BE]: pyupgrade Python to 3.8 - imports and object inheritance only (#94308) 2023-02-07 21:10:56 +00:00
rendezvous.py Revisit torch._six.string_classes removal (#94709) (#97863) 2023-03-30 17:02:45 +00:00
run.py Convert logging f-strings to use % format, part four (#98705) 2023-04-11 13:17:59 +00:00
utils.py [BE]: Bugfix functorch and some generic typing improvements (#101337) 2023-05-14 14:20:56 +00:00