FP16 fused and unfused grad norm query. #1286
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
ShadenSmith
requested review from
RezaYazdaniAminabadi,
awan-10,
cli99,
conglongli,
eltonzheng,
jeffra,
minjiaz,
niumanar,
samyam and
tjruwase
as code owners
Contributor
Author
|
This still needs FP32 and ZeRO. And unit tests :-). |
stas00
reviewed
Aug 7, 2021
| self._config.gradient_accumulation_steps = new_gas | ||
|
|
||
| def _compute_global_grad_norm(self): | ||
| params = [p for p in self.module.parameters() if p.grad is not None] |
Collaborator
There was a problem hiding this comment.
would it be a bit more efficient if tied params don't get calculated more than once? Probably, something like:
params = dict((p.data_ptr(), p) for p in self.module.parameters() if p.grad is not None).values()
or it might help to have a wrapper to do that, since this is a handy util.
But please verify that I got it right. Thanks.
|
Can one of the admins verify this patch? |
Collaborator
|
This ended up being added in a later PR |
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Outdated suggestions cannot be applied.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.
No description provided.