[FSDP][state_dict] Restore the state_dict_config for NO_SHARD #100855
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Any change to the user configurations should be temporary. This PR fixes the issue when NO_SHARD state_dict/load_state_dict is called, the state_dict_config and state_dict_type are changed permanently. Differential Revision: [D45593313](https://our.internmc.facebook.com/intern/diff/D45593313/) **NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D45593313/)! [ghstack-poisoned]
fegin
requested review from
mrshenli,
zhaojuanmao,
rohan-varma,
H-Huang,
awgu,
kwen2501,
wanchaol,
kiukchung and
d4l3k
as code owners
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/100855
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 4d43a7d: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
pytorch-bot
bot
added
the
release notes: distributed (fsdp)
Any change to the user configurations should be temporary. This PR fixes the issue when NO_SHARD state_dict/load_state_dict is called, the state_dict_config and state_dict_type are changed permanently. Differential Revision: [D45593313](https://our.internmc.facebook.com/intern/diff/D45593313/) **NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D45593313/)! [ghstack-poisoned]
fegin
changed the title
…ARD" Any change to the user configurations should be temporary. This PR fixes the issue when NO_SHARD state_dict/load_state_dict is called, the state_dict_config and state_dict_type are changed permanently. Differential Revision: [D45593313](https://our.internmc.facebook.com/intern/diff/D45593313/) **NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D45593313/)! [ghstack-poisoned]
awgu
approved these changes
May 9, 2023
| @@ -302,6 +302,8 @@ def test_state_dict_fsdp_submodules(self): | |||
| self.assertIsInstance(tensor, ShardedTensor) | |||
| elif "u2" in fqn: | |||
| self.assertIsInstance(tensor, torch.Tensor) | |||
| # Ensure that get_state_dict_type can still correctly get the settings. | |||
| _ = FSDP.get_state_dict_type(model) | |||
There was a problem hiding this comment.
Is this meant to have some assert on the return value?
There was a problem hiding this comment.
get_state_dict_type will assert if the state_dict_config is not restored.
zhaojuanmao
approved these changes
May 9, 2023
…ARD" Any change to the user configurations should be temporary. This PR fixes the issue when NO_SHARD state_dict/load_state_dict is called, the state_dict_config and state_dict_type are changed permanently. Differential Revision: [D45593313](https://our.internmc.facebook.com/intern/diff/D45593313/) **NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D45593313/)! [ghstack-poisoned]
Merge failedReason: Comment with id 1540563109 not found Details for Dev Infra teamRaised by workflow job |
…ARD" Any change to the user configurations should be temporary. This PR fixes the issue when NO_SHARD state_dict/load_state_dict is called, the state_dict_config and state_dict_type are changed permanently. Differential Revision: [D45593313](https://our.internmc.facebook.com/intern/diff/D45593313/) **NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D45593313/)! [ghstack-poisoned]
fegin
added a commit
that referenced
this pull request
May 9, 2023
Pull Request resolved: #100855 Any change to the user configurations should be temporary. This PR fixes the issue when NO_SHARD state_dict/load_state_dict is called, the state_dict_config and state_dict_type are changed permanently. ghstack-source-id: 188583060 Differential Revision: [D45593313](https://our.internmc.facebook.com/intern/diff/D45593313/) **NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D45593313/)!
…ARD" Any change to the user configurations should be temporary. This PR fixes the issue when NO_SHARD state_dict/load_state_dict is called, the state_dict_config and state_dict_type are changed permanently. Differential Revision: [D45593313](https://our.internmc.facebook.com/intern/diff/D45593313/) **NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D45593313/)! [ghstack-poisoned]
rohan-varma
approved these changes
May 10, 2023
|
@pytorchbot merge (Initiating merge automatically since Phabricator Diff has merged) |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Stack from ghstack (oldest at bottom):
Any change to the user configurations should be temporary. This PR fixes the issue when NO_SHARD state_dict/load_state_dict is called, the state_dict_config and state_dict_type are changed permanently.
Differential Revision: D45593313
NOTE FOR REVIEWERS: This PR has internal Meta-specific changes or comments, please review them on Phabricator!