[Model]: add FLUX.2-dev model

[nuclearwu]

PLEASE FILL IN THE PR DESCRIPTION HERE ENSURING ALL CHECKLIST ITEMS (AT THE BOTTOM) HAVE BEEN CONSIDERED.

Purpose

support https://huggingface.co/black-forest-labs/FLUX.2-dev

Test Plan

vLLM-Omni:

python examples/offline_inference/text_to_image/text_to_image.py \
  --model /workspace/cache/ymttest/johnjan/models/black-forest-labs/FLUX___2-dev/ \
  --prompt "a lovely bunny holding a sign that says 'vllm-omni'" \
  --seed 42 \
  --tensor-parallel-size 2 \
  --num-images-per-prompt 1 \
  --num-inference-steps 50 \
  --guidance-scale 4.0 \
  --height 1024 \
  --width 1024 \
  --output outputs/flux2-dev.png

Test Result

vLLM-Omni:
Reproduced with 4xA800.

Model/TP	diffusers	TP=1	TP=2	TP=4
Flux.2-dev		OOM
Time	104.9411s/img	OOM	39.1087s/img	29.0770s/img

---

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan. Please provide the test scripts & test commands. Please state the reasons if your codes don't require additional test scripts. For test file guidelines, please check the test style doc
• The test results. Please paste the results comparison before and after, or the e2e results.
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model. Please run mkdocs serve to sync the documentation editions to ./docs.
• (Optional) Release notes update. If your change is user-facing, please update the release notes draft.

BEFORE SUBMITTING, PLEASE READ https://github.com/vllm-project/vllm-omni/blob/main/CONTRIBUTING.md (anything written below this line will be removed by GitHub Actions)

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a53145a246

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Allow embedding-only prompts by preserving None prompt

When a request supplies prompt_embeds without a text prompt, this line coerces missing text to "" instead of keeping it None. That makes check_inputs think both prompt and prompt_embeds were provided and raises a ValueError, so embedding-only inference for Flux2Pipeline is effectively broken.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Invoke step-end callback during denoising iterations

The pipeline accepts callback_on_step_end and validates callback_on_step_end_tensor_inputs, but the denoising loop never invokes the callback. Any caller relying on per-step hooks (progress reporting, custom stopping, or latent inspection) will silently lose that behavior in Flux2Pipeline compared with other diffusion pipelines in this repo.

Useful? React with 👍 / 👎.

[mergify]

⚠️ The sha of the head commit of this PR conflicts with #1629. Mergify cannot evaluate rules on this PR. Once #1629 is merged or closed, Mergify will resume processing this PR. ⚠️