Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Add Krea 2 (K2) text-to-image pipeline and transformer#14045

Merged
asomoza merged 1 commit into
mainfrom
krea2
Jun 22, 2026
Merged

Add Krea 2 (K2) text-to-image pipeline and transformer#14045
asomoza merged 1 commit into
mainfrom
krea2

Conversation

@yiyixuxu

Copy link
Copy Markdown
Collaborator

No description provided.

Adds Krea2Transformer2DModel and Krea2Pipeline: a flow-matching
text-to-image MMDiT with grouped-query attention, Qwen3-VL text
conditioning, and the Qwen-Image VAE.

Co-authored-by: Naga Sai Abhinay Devarinti <[email protected]>
@github-actions github-actions Bot added documentation Improvements or additions to documentation models tests utils size/L PR with diff > 200 LOC pipelines labels Jun 22, 2026
@HuggingFaceDocBuilderDev

Copy link
Copy Markdown

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

@asomoza asomoza left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thanks!

@asomoza asomoza merged commit 7104cb4 into main Jun 22, 2026
16 checks passed
MichaelRicks pushed a commit to MichaelRicks/LTX-Desktop-Mike that referenced this pull request Jun 30, 2026
Krea2Pipeline/Krea2Transformer2DModel only exist on diffusers main as of the
2026-06-22 merge (huggingface/diffusers#14045) — diffusers' latest PyPI
release still predates it. Pin to that exact merge commit (not a moving
"main" pointer) to keep the dependency surface minimal. safetensors came
along as a transitive bump; nothing else moved.

Verified: full backend test suite passes (257 tests), and confirmed live
that existing Z-Image-Turbo image generation and LTX video generation both
still work after the bump, before writing any Krea 2 specific code.
DN6 pushed a commit that referenced this pull request Jul 1, 2026
Adds Krea2Transformer2DModel and Krea2Pipeline: a flow-matching
text-to-image MMDiT with grouped-query attention, Qwen3-VL text
conditioning, and the Qwen-Image VAE.

Co-authored-by: Elea Zhong <[email protected]>
Co-authored-by: Naga Sai Abhinay Devarinti <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

documentation Improvements or additions to documentation models pipelines size/L PR with diff > 200 LOC tests utils

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants