Fix PT TF ViTMAE #16766
Merged
Fix PT TF ViTMAE #16766
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
ydshieh
requested review from
NielsRogge,
gante and
sgugger
and removed request for
sgugger
|
Let's merge since CI fails quite a lot on this one |
Ok! |
elusenji
pushed a commit
to elusenji/transformers
that referenced
this pull request
Jun 12, 2022
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.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.
What does this PR do?
Fix PT TF ViTMAE: just use some settings both in PT/TF (instead of in only one model). Otherwise, the PT/TF equivalence tests for them won't use something like
std = 0.02, and gets larger (init) weights --> larger diff in outputs.Also, the
epsforlayer normlayers should be the same in PT/TF.(not a real big deal in practice, since here is
1e-5v.s.1e-12-> but it also affects the tests)