XLNet Bug when training with apex 16-bit precision #6567
Merged
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
Codecov Report
@@ Coverage Diff @@
## master #6567 +/- ##
==========================================
- Coverage 79.18% 78.41% -0.78%
==========================================
Files 156 156
Lines 28129 28129
==========================================
- Hits 22275 22056 -219
- Misses 5854 6073 +219
Continue to review full report at Codecov.
|
JetRunner
reviewed
Aug 20, 2020
|
Of course, @JetRunner, here it is |
JetRunner
approved these changes
Aug 20, 2020
There was a problem hiding this comment.
Of course, @JetRunner, here it is
Great! Would you mind adding a one-line comment explaining the cast?
|
@JetRunner, done |
|
Merging since the CI error looks unrelated. |
Zigur
pushed a commit
to Zigur/transformers
that referenced
this pull request
Oct 26, 2020
* xlnet fp16 bug fix * comment cast added * Update modeling_xlnet.py Co-authored-by: Kevin Canwen Xu <canwenxu@126.com>
fabiocapsouza
pushed a commit
to fabiocapsouza/transformers
that referenced
this pull request
Nov 15, 2020
* xlnet fp16 bug fix * comment cast added * Update modeling_xlnet.py Co-authored-by: Kevin Canwen Xu <canwenxu@126.com>
fabiocapsouza
added a commit
to fabiocapsouza/transformers
that referenced
this pull request
Nov 15, 2020
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Outdated suggestions cannot be applied.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.
XLNet training fail, while using 16-bit precision, because of tensor creation with explicit usage dtype=torch.float mode in relative_positional_encodings function.