Transformer and sampling implementation adapted from:
https://github.com/black-forest-labs/flux-3-image-klein
Revision: 45e57c9fb06f9751cdced29a65c5d2171003be05
Copyright Black Forest Labs. Apache License 2.0; see LICENSE.

Sampling shift updated from upstream revision:
62e90b9278b73ecc52927770c6965dc6b1966498

Enhancer instructions condensed and adapted from src/flux3/pup_prompts_external_vlm.json:
https://github.com/black-forest-labs/flux-3-image-klein
Revision: 458dcc00e4719f65be0d9a611d52d9cf86ba980d
Copyright Black Forest Labs. Apache License 2.0; see LICENSE.
Adaptations preserve the user's language, support requested display-copy authoring,
and use WanGP's generation-context placeholders.

Qwen3-VL components are reused from WanGP's local Transformers implementation.
Copyright 2025 The Qwen Team and The HuggingFace Inc. team; Apache License 2.0.

Model weights have separate terms; see WEIGHTS_LICENSE.md and WEIGHTS_NOTICE.md.
