fix(z_image): use unrestricted image self-attention for regional prompting#8718
Merged
lstein merged 1 commit intoinvoke-ai:mainfrom Dec 28, 2025
Conversation
…pting Changes image self-attention from restricted (region-isolated) to unrestricted (all image tokens can attend to each other), similar to the FLUX approach. This fixes the issue where ZImage-Turbo with multiple regional guidance layers would generate two separate/disconnected images instead of compositing them into a single unified image. The regional text-image attention remains restricted so that each region still responds to its corresponding prompt. Fixes invoke-ai#8715
Pfannkuchensack
requested review from
blessedcoolant and
lstein
as code owners
github-actions
bot
added
python
1 task
lstein
approved these changes
Dec 28, 2025
Collaborator
lstein
left a comment
lstein
left a comment
There was a problem hiding this comment.
This change works well in my hands. I find that even minimal overlap of the regions ensures better coherence to the main prompt, but I usually get acceptable results even when the regions are not overlapping.
Pfannkuchensack
deleted the
fix/zimage-regional-guidance-unrestricted-attention
branch
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Changes image self-attention from restricted (region-isolated) to unrestricted (all image tokens can attend to each other), similar to the FLUX approach.
This fixes the issue where ZImage-Turbo with multiple regional guidance layers would generate two separate/disconnected images instead of compositing them into a single unified image.
The regional text-image attention remains restricted so that each region still responds to its corresponding prompt.
Fixes #8715
Related Issues / Discussions
QA Instructions
Merge Plan
No special merge considerations required. This is a targeted fix to the attention mask construction.
Checklist
What's Newcopy (if doing a release after this PR)