Fix LCM LoRA SDXL distillation state dict loading and saving - #14562
Open
loulanyue wants to merge 1 commit into
Open
Fix LCM LoRA SDXL distillation state dict loading and saving#14562loulanyue wants to merge 1 commit into
loulanyue wants to merge 1 commit into
Conversation
Keep native PEFT state dict without routing through convert_state_dict_to_diffusers, ensuring both intermediate validation and saved checkpoints preserve all attention and convolutional adapters when loaded into StableDiffusionXLPipeline. Fixes huggingface#14548
github-actions
Bot
added
fixes-issue
size/S
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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?
Fixes #14548
In
examples/consistency_distillation/train_lcm_distill_lora_sdxl.py, routing the UNet's PEFT state dict throughconvert_state_dict_to_diffuserscauses two issues:StableDiffusionXLPipeline.load_lora_weights, causing it to drop the weights and log un-adapted base model outputs.convert_state_dict_to_diffusersonly renames attention modules to legacy naming while leaving convolutional and projection modules in PEFT naming, resulting in dropped attention adapters when loaded downstream.This PR removes the unnecessary
convert_state_dict_to_diffusersconversion intrain_lcm_distill_lora_sdxl.py, keeping PEFT-native keys consistent across saving and loading hooks.Who can review?
@sayakpaul @linoytsaban