Fix LayerScale ignoring init_values#2605
Merged
rwightman merged 1 commit intohuggingface:mainfrom Nov 4, 2025
Merged
Conversation
Collaborator
|
@Ilya-Fradlin thanks, FYI I have quite a bit set of changes like/related to this to implement a two phase init, but will merge this one regardless. EDIT woops I cut that off... becuase there's a regression here, I removed a number of local 'LayerScale' impl that did not have this issue and used this central one which does have the problematic reset... was probably in my changes a month back. |
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Previously, reset_parameters() set gamma to ones, ignoring init_values and effectively disabling LayerScale’s small-init behavior.
Now, gamma is initialized to init_values via nn.init.constant_ preserving the intended effect.