🚨🚨🚨 Remove softmax for EfficientNetForImageClassification 🚨🚨🚨#25501
Merged
amyeroberts merged 3 commits intohuggingface:mainfrom Aug 14, 2023
Merged
🚨🚨🚨 Remove softmax for EfficientNetForImageClassification 🚨🚨🚨#25501amyeroberts merged 3 commits intohuggingface:mainfrom
amyeroberts merged 3 commits intohuggingface:mainfrom
Conversation
Contributor
Author
|
cc @rwightman Managed to track down the issue with the efficient net predictions on the hub |
|
The documentation is not available anymore as the PR was closed or merged. |
sgugger
approved these changes
Aug 14, 2023
Collaborator
sgugger
left a comment
There was a problem hiding this comment.
Nice catch, and necessary breaking change, I agree!
5 tasks
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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?
/!\ This is a breaking change /!\
EfficientNet implementation erroneously added a softmax to the models logits in the classification head. This breaks with convention and documentation for the model outputs.
This results in the very small probabilities seen in the hosted widgets on the checkpoints pages, as the logits effectively have softmax applied twice.
Most of the efficientnet checkpoints were downloaded < 100 times in the past month. However, efficientnet-b7 is more popular with a few thousand downloads.
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.