Fix tf shared embedding #17730
Merged
Fix tf shared embedding #17730
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
|
All the TF weights of OPT will need to be updated if this is approved. I think I can handle that along with #17713. |
sgugger
approved these changes
Jun 16, 2022
| @@ -281,14 +281,14 @@ def test_inference_no_head(self): | |||
|
|
|||
|
|
|||
| @require_tf | |||
| @slow | |||
| # @slow | |||
There was a problem hiding this comment.
Will need to be uncommented before merging.
There was a problem hiding this comment.
Think it's already fixed in the new commit
patrickvonplaten
approved these changes
Jun 16, 2022
|
The documentation is not available anymore as the PR was closed or merged. |
sgugger
pushed a commit
that referenced
this pull request
Jun 16, 2022
* fix the naming * from pt in test for now * make style * slow test and removed from_pt
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Outdated suggestions cannot be applied.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.
What does this PR do?
A hack was used to properly import the shared embedding weights, but it can be removed (removing it also is convenient for the sharding PR)
Found this while testing #17713. In HF's
save_pretrainedandload_pretrainedthe layer name is changed usingname = "/".join(weight_name.split("/")[1:]). This was breaking for OPT as the layer name was'decoder.embed_tokens/model.decoder.embed_tokens/weight:0'instead of'tfopt_model/model/decoder/embed_tokens/weight:0'. The naming is strange and had to use a scope hack. The hack comes fromBART