[LLM] [Cherry-Pick] valid loss before optimizer step (#9255) #9705
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
|
Thanks for your contribution! |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## develop #9705 +/- ##
===========================================
+ Coverage 51.99% 53.02% +1.02%
===========================================
Files 726 718 -8
Lines 115607 112374 -3233
===========================================
- Hits 60110 59581 -529
+ Misses 55497 52793 -2704 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
wawltor
reviewed
Dec 27, 2024
| @@ -1133,6 +1133,9 @@ def fused_allreduce_gradients_no_sync(paramlist, hcg): | |||
| if self.args.pipeline_parallel_degree <= 1 and self._enable_delay_scale_loss(): | |||
| tr_loss /= self.args.gradient_accumulation_steps | |||
|
|
|||
| # assert if loss is invalid | |||
| self._check_loss_valid(tr_loss) | |||
|
|
|||
There was a problem hiding this comment.
有个疑问，如果是float16的训练出现nan或者inf，会直接报错？
There was a problem hiding this comment.
fp16训练不会报错，在下面的函数实现里会有判断。
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Outdated suggestions cannot be applied.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.
PR types
New features
PR changes
Others
Description
CP from: #9255
report error when getting loss of NaN of Inf, for LLM training platform integration.
[Pcard-88789]