add cache-dit support for Qwen-Image #61
Merged
+3
−0
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
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Outdated suggestions cannot be applied.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.
cache-dit
cache-dit offers cache acceleration support for Qwen-Image with DBCache, TaylorSeer and Cache CFG. Visit their example for more details.
Baseline
cache-dit: DBCache F8B0 + TaylorSeer O(2) + Cache CFG, 1.5x Speed Up
cache-dit: DBCache F8B0 + TaylorSeer O(4) + Cache CFG, 1.5x Speed Up
We can see that the higher the TaylorSeer order, the better the precision. The image of TaylorSeer O(4) generates content above the small blackboard that is basically consistent with the Baseline, while TaylorSeer O(2) will generate a new surround light tube.