fix(controlnet): Use deep copy in ZImageControlNet.from_transformer#13102
Open
Mr-Neutr0n wants to merge 1 commit intohuggingface:mainfrom
Open
fix(controlnet): Use deep copy in ZImageControlNet.from_transformer#13102Mr-Neutr0n wants to merge 1 commit intohuggingface:mainfrom
Mr-Neutr0n wants to merge 1 commit intohuggingface:mainfrom
Conversation
The from_transformer classmethod was creating shallow copies of modules from the transformer, causing modifications to the controlnet weights to also affect the original transformer weights. This fix uses copy.deepcopy() to ensure the controlnet has its own independent copy of the weights. Fixes huggingface#13077
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Use
copy.deepcopy()instead of direct assignment inZImageControlNet.from_transformer()to prevent weight sharing between controlnet and transformer.Problem
The
from_transformermethod was using direct assignment to copy modules from transformer to controlnet. This creates a shallow copy where both objects share the same underlying tensor references. Training the controlnet would inadvertently modify the original transformer weights.Solution
Changed all module assignments to use
copy.deepcopy():t_embedderall_x_embeddercap_embedderrope_embeddernoise_refinercontext_refinerx_pad_tokencap_pad_tokenNote:
t_scaleis a scalar value (not a module), so direct assignment is correct for it.Fixes #13077