Chien-Chin Huang 8c57fde21f Let all_reduce_coalesced accept one tensor as well (#115650) ... This diff introduces a change to the `all_reduce_coalesced` function in `distributed_c10d.py`. The function now accepts a single tensor as well as a list of tensors. This allows for more flexibility in the use of the function. This is just a syntax sugar for the compiler to use `all_reduce_coalesced` without worrying about converting the input to a list. Differential Revision: [D51433236](https://our.internmc.facebook.com/intern/diff/D51433236/) Pull Request resolved: https://github.com/pytorch/pytorch/pull/115650 Approved by: https://github.com/wconstab ghstack dependencies: #115523, #115302, #115648, #115649		2023-12-13 21:32:01 +00:00
..
_composable
_shard	Updated docs for deprecated torch.set_default_tensor_type (#115041)	2023-12-07 16:17:36 +00:00
_sharded_tensor
_sharding_spec
_spmd
_tensor	Adds tool to visualize sharding (#114307)	2023-12-12 06:18:03 +00:00
_tools
algorithms
autograd
benchmarks
checkpoint	[DCP][BE] Apply ufmt to DCP and turn on lintrunner for DCP (#115302)	2023-12-13 10:32:36 +00:00
elastic	Switch env variable use in test harnesses to the non-deprecated names to fix warnings (#114880)	2023-12-01 20:08:23 +00:00
examples
fsdp	[fsdp] Replace acc_grad hooking with register_post_accumulate_grad_hook on flat_param (#112184)	2023-12-13 16:24:44 +00:00
launcher
nn
optim
pipeline
rpc
tensor	[DeviceMesh] Rename _device_mesh.py to device_mesh.py to prepare for beta (#115099) (#115193)	2023-12-08 08:44:32 +00:00
__init__.py	Fix torch.distributed.breakpoint (#115705)	2023-12-13 20:33:56 +00:00
_composable_state.py
_functional_collectives.py	[funcol] Directly import DeviceMesh to avoid circular dependency (#115649)	2023-12-13 20:44:58 +00:00
_functional_collectives_impl.py	[funcol][BE] Apply ufmt to _functional_collectives.py and turn on lintrunner for functional_collective (#115648)	2023-12-13 11:19:29 +00:00
_state_dict_utils.py	[DCP][BE] Move DCP._state_dict_utils out from DCP (#115523)	2023-12-13 08:59:48 +00:00
argparse_util.py
c10d_logger.py	Re-enable type checking for distributed_c10d.py (#115223)	2023-12-09 11:07:54 +00:00
collective_utils.py
constants.py	Switch env variable use in test harnesses to the non-deprecated names to fix warnings (#114880)	2023-12-01 20:08:23 +00:00
CONTRIBUTING.md
device_mesh.py	[DeviceMesh] Move helper function 'get_mesh_dim_by_name' to MeshEnv class (#115572)	2023-12-12 06:29:46 +00:00
distributed_c10d.py	Let all_reduce_coalesced accept one tensor as well (#115650)	2023-12-13 21:32:01 +00:00
launch.py
logging_handlers.py
remote_device.py
rendezvous.py
run.py
utils.py