[PT-D][FSDP] Implement _clip_grad_norm_ for FSDP #73405
Closed
Conversation
This file contains hidden or 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
Implement the `_clip_grad_norm_` for FSDP, issue: #72548 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/) [ghstack-poisoned]
fduwjj
requested review from
H-Huang,
mingzhe09088,
mrshenli,
pritamdamania87,
rohan-varma and
zhaojuanmao
as code owners
Contributor
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit a87e2f9 (more details on the Dr. CI page): 💚 💚 Looks good so far! There are no failures yet. 💚 💚 This comment was automatically generated by Dr. CI (expand for details).Please report bugs/suggestions to the (internal) Dr. CI Users group. |
CI Flow Status⚛️ CI FlowRuleset - Version:
|
facebook-github-bot
added
cla signed
oncall: distributed
fduwjj
added a commit
that referenced
this pull request
Feb 25, 2022
Implement the `_clip_grad_norm_` for FSDP, issue: #72548 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/) ghstack-source-id: 149925805 Pull Request resolved: #73405
rohan-varma
reviewed
Feb 25, 2022
Contributor
rohan-varma
left a comment
rohan-varma
left a comment
There was a problem hiding this comment.
This is awesome work, thanks so much for persisting through all the usability issues and giving feedback on them offline! Added some comments for your consideration.
zhaojuanmao
reviewed
Mar 1, 2022
Contributor
zhaojuanmao
left a comment
zhaojuanmao
left a comment
There was a problem hiding this comment.
nice work! thanks for adding this.
for tests, would you please add a norm_type config to _train_for_several_steps() in common_fsdp.py and compare ddp vs fsdp parity, similar to fairscale tests in _train_for_several_steps of test_fsdp.py for test_clip_norm_transformer and test_mixture_of_experts_grad_clip_breaks
Implement the `_clip_grad_norm_` for FSDP, issue: #72548 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/) [ghstack-poisoned]
Implement the `_clip_grad_norm_` for FSDP, issue: #72548 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/) [ghstack-poisoned]
fduwjj
added a commit
that referenced
this pull request
Mar 7, 2022
Pull Request resolved: #73405 Implement the `_clip_grad_norm_` for FSDP, issue: #72548 ghstack-source-id: 150655449 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/)
Contributor
Author
|
Addressed the comment from reviewers. |
rohan-varma
reviewed
Mar 8, 2022
rohan-varma
reviewed
Mar 8, 2022
rohan-varma
reviewed
Mar 8, 2022
Implement the `_clip_grad_norm_` for FSDP, issue: #72548 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/) [ghstack-poisoned]
fduwjj
added a commit
that referenced
this pull request
Mar 9, 2022
Pull Request resolved: #73405 Implement the `_clip_grad_norm_` for FSDP, issue: #72548 ghstack-source-id: 150951935 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/)
Contributor
Author
|
Adding the test coverage for the nested model. |
zhaojuanmao
reviewed
Mar 9, 2022
Contributor
zhaojuanmao
left a comment
zhaojuanmao
left a comment
There was a problem hiding this comment.
overall looks great to me, just one minor comment, will let @rohan-varma to accept it.
Implement the `_clip_grad_norm_` for FSDP, issue: #72548 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/) [ghstack-poisoned]
fduwjj
added a commit
that referenced
this pull request
Mar 10, 2022
Pull Request resolved: #73405 Implement the `_clip_grad_norm_` for FSDP, issue: #72548 ghstack-source-id: 150990293 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/)
Contributor
Author
|
Further address reviewer's comment. |
rohan-varma
reviewed
Mar 10, 2022
rohan-varma
approved these changes
Mar 10, 2022
Contributor
rohan-varma
left a comment
rohan-varma
left a comment
There was a problem hiding this comment.
Awesome work, thanks for persisting through all of the issues and addressing all the comments! Looks great to ship.
Implement the `_clip_grad_norm_` for FSDP, issue: #72548 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/) [ghstack-poisoned]
fduwjj
added a commit
that referenced
this pull request
Mar 10, 2022
Pull Request resolved: #73405 Implement the `_clip_grad_norm_` for FSDP, issue: #72548 ghstack-source-id: 151059433 Differential Revision: [D34230605](https://our.internmc.facebook.com/intern/diff/D34230605/)
Contributor
|
Hey @fduwjj. |
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):
Implement the
_clip_grad_norm_for FSDP, issue: #72548Differential Revision: D34230605