[WIP] Add Flux2 modular

[DN6]

What does this PR do?

Fixes # (issue)

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline?
• Did you read our philosophy doc (important for complex PRs)?
• Was this discussed/approved via a GitHub issue or the forum? Please add a link to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

[HuggingFaceDocBuilderDev]

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.

[yiyixuxu]

thanks! I left some feedbacks, but feel free to merge once it's working:)

[yiyixuxu]

it's up to you, but I think maybe we don't need to mix the remote blocks into Auto - we can separately list them so people can mix it on their own

ALL_BLOCKS = {
    "text2image": TEXT2IMAGE_BLOCKS,
    "image_conditioned": IMAGE_CONDITIONED_BLOCKS,
    "auto": AUTO_BLOCKS,
    "remote" : 
}

[yiyixuxu]

ohh, I think we don't need to wrap it inside a AutoPipelineBlocks if this is the only block

[yiyixuxu]

same here

[yiyixuxu]

same here too, we don't need to wrap it inside an autoblock
if the decode step works for all workflow we esupport,

[yiyixuxu]

so this is the set we use for UI, so basically we need to have these 4 blocks

• text_encoder
• vae_image_encoder
• denoise
• decode

each corresponding to a node

I would pack text_input/image_input/before_Denoise/denoise into an "AutoDenoise" block, you can pack it however you like, but I think one way to do it is to:

first, have 2 sequential blocks
for text-to-image: text_input -> before_denoise -> denoise
for image condition: text_input -> image_input -> before_denoise -> denoise

and then pack these two into an AutoDenoise

[yiyixuxu]

Suggested change

	("image_encoder", Flux2AutoVaeEncoderStep()),
	("vae_image_encoder", Flux2AutoVaeEncoderStep()),

naming it "vae_image_encoder" because there are also image_encoder blocks using like clip/siglip image encoder, just