Optionally bundle joint-trained embeddings into LoRAs #278
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
Not sure if anything respects the format for these, but it matches SD/SDXL which is implemented in A1111.
Nerogar
reviewed
May 9, 2024
This was referenced Jun 5, 2024
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Outdated suggestions cannot be applied.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.
This uses the format defined by AUTOMATIC1111/stable-diffusion-webui#13568. I tested it for SD1.5; I can't run any of the others on my machine so did not test those.
Currently A1111 only supports SD1.5/SD2/SDXL. There's no Wuerstchen or Pixart support in anything to actually use the bundled embeddings. That said, there was no reason to not extend the format to them.
I only write out the embeddings under
model.additional_embeddingsrather than anything inmodel.additional_embedding_states, with the idea that just because an embedding is present in training doesn't mean it should be shipped with the model. If I've misunderstood something about howmodel.additional_embedding_statesis used, please correct me on this.