HiDreams w/ llama only works better with cached encoded empty string conditionings than with all zeroes. #7701
+17
−10
Conversation
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
…lity and prompt adherence with llama-only HiDreams
|
Can you please explain what to change in code to use your test cached encoded empty files. I downloaded them but don't know what to change in hidream.py. Thx |
|
Thank you very much for pointing it out for me. I was not able to find it. |
|
it works, if the devs include it , this will be handy for weaker gpus. first one is quad clip second one is this implementation |
Nice...I re-ran your workflow with Llama only, HiDream Full uncensored, using my new favourite sampler Distance+Beta, in only 10 steps (73.64sec): Compared to UniPC+Simple, in 30 steps (79.21sec): |
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Outdated suggestions cannot be applied.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 is similar to my previous pull request, but it should work fine with the changes the other PR added. What it didn't add were cached, encoded empty prompts for the pooled and t5 outputs, and instead it continued the behavior of sending all zeroes for those.
Llama only, with torch.zeroes for other encoders:
Llama only, with cached encoded empty strings for other encoders:
