[FSDP] Ensure that customized non tensor optimizer state can be saved #99214
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
The current logic does not actually handle all different non-tensor optimizer states correctly. This PR fixes the issue and adds a test. Differential Revision: [D45021331](https://our.internmc.facebook.com/intern/diff/D45021331/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/99214
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit fd264bd: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
fegin
requested review from
mrshenli,
zhaojuanmao and
rohan-varma
as code owners
pytorch-bot
bot
added
the
release notes: distributed (fsdp)
fegin
requested review from
H-Huang,
awgu,
kwen2501,
wanchaol,
kiukchung and
d4l3k
as code owners
fegin
added a commit
that referenced
this pull request
Apr 15, 2023
The current logic does not actually handle all different non-tensor optimizer states correctly. This PR fixes the issue and adds a test. Differential Revision: [D45021331](https://our.internmc.facebook.com/intern/diff/D45021331/) ghstack-source-id: 186204536 Pull Request resolved: #99214
awgu
approved these changes
Apr 15, 2023
| osd = FSDP.optim_state_dict(model, optim, optim_state_dict=original_osd) | ||
| osd_to_load = FSDP.optim_state_dict_to_load(model, optim, osd ) | ||
| for param_id, state in osd_to_load["state"].items(): | ||
| # Addd customized value |
There was a problem hiding this comment.
nit: typo 😄
Suggested change
| # Addd customized value | |
| # Add customized value |
| step() | ||
| original_osd = deepcopy(optim.state_dict()) | ||
| for param_id, state in original_osd["state"].items(): | ||
| # Addd customized value |
There was a problem hiding this comment.
Suggested change
| # Addd customized value | |
| # Add customized value |
| state["value2"] = None | ||
|
|
||
| osd = FSDP.optim_state_dict(model, optim, optim_state_dict=original_osd) | ||
| osd_to_load = FSDP.optim_state_dict_to_load(model, optim, osd ) |
There was a problem hiding this comment.
nit: I wonder if lint will complain about this one
Suggested change
| osd_to_load = FSDP.optim_state_dict_to_load(model, optim, osd ) | |
| osd_to_load = FSDP.optim_state_dict_to_load(model, optim, osd) |
…an be saved" The current logic does not actually handle all different non-tensor optimizer states correctly. This PR fixes the issue and adds a test. Differential Revision: [D45021331](https://our.internmc.facebook.com/intern/diff/D45021331/) [ghstack-poisoned]
fegin
added a commit
that referenced
this pull request
Apr 15, 2023
Pull Request resolved: #99214 The current logic does not actually handle all different non-tensor optimizer states correctly. This PR fixes the issue and adds a test. ghstack-source-id: 186225950 Differential Revision: [D45021331](https://our.internmc.facebook.com/intern/diff/D45021331/)
…an be saved" The current logic does not actually handle all different non-tensor optimizer states correctly. This PR fixes the issue and adds a test. This PR will solve #99079 Differential Revision: [D45021331](https://our.internmc.facebook.com/intern/diff/D45021331/) [ghstack-poisoned]
fegin
added a commit
that referenced
this pull request
Apr 16, 2023
Pull Request resolved: #99214 The current logic does not actually handle all different non-tensor optimizer states correctly. This PR fixes the issue and adds a test. ghstack-source-id: 186239769 Differential Revision: [D45021331](https://our.internmc.facebook.com/intern/diff/D45021331/)
|
@pytorchbot merge |
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 |
Merge failedReason: 1 jobs have failed, first few of them are: trunk / macos-12-py3-arm64 / build Details for Dev Infra teamRaised by workflow job |
|
@pytorchbot merge |
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 |
4 tasks
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):
The current logic does not actually handle all different non-tensor optimizer states correctly. This PR fixes the issue and adds a test.
This PR will solve #99079
Differential Revision: D45021331