Skip to content

[hotfix] shared model returns cpu state_dict #1328

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
feifeibear merged 2 commits into from
Jul 15, 2022
Merged

[hotfix] shared model returns cpu state_dict #1328

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3a12b13
shared model returns cpu state_dict
ver217 Jul 15, 2022
c40e70d
update unit test
ver217 Jul 15, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion colossalai/zero/sharded_model/sharded_model_v2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -439,7 +439,8 @@ def _colo_state_dict(self,
for p in sharded_params:
p.data = p.colo_attr.data_payload
module_to_load = module_to_load or self
gathered_state_dict = deepcopy(state_dict_func(module_to_load, destination, prefix, keep_vars))
gathered_state_dict = state_dict_func(module_to_load, destination, prefix, keep_vars)
gathered_state_dict = {k: v.cpu() if isinstance(v, torch.Tensor) else v for k, v in gathered_state_dict.items()}
if shard_strategy is not None:
shard_strategy.shard([p.colo_attr.sharded_data_tensor for p in sharded_params], process_group)
for p in sharded_params:
Expand Down
2 changes: 1 addition & 1 deletion tests/test_zero/test_state_dict.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,7 +39,7 @@ def run_zero_state_dict(shard_strategy_class):

zero_state_dict = zero_model.state_dict()
for key, val in model.state_dict().items():
assert torch.equal(val, zero_state_dict[key])
assert torch.equal(val, zero_state_dict[key].to(val.device))


def run_dist(rank, world_size, port):
Expand Down