Align Flux modular more and more with Qwen modular

[sayakpaul]

What does this PR do?

As the title goes.

Next plan is to incorporate these changes in #12269.

I tested with the code snippet from #12419 (comment).

For img2img, I did:

import torch 
from diffusers import ModularPipeline
from diffusers.utils import load_image

repo_id = "black-forest-labs/FLUX.1-dev"

pipe = ModularPipeline.from_pretrained(repo_id)
pipe.load_components(torch_dtype=torch.bfloat16)
pipe = pipe.to("cuda")
# print(pipe)

url = "https://raw.githubusercontent.com/CompVis/stable-diffusion/main/assets/stable-samples/img2img/sketch-mountains-input.jpg"
init_image = load_image(url).resize((1024, 1024))

output = pipe(
    prompt="a dog sitting by the see waiting for its companion to come",
    image=init_image,
    guidance_scale=3.5,
    num_inference_steps=28,
    max_sequence_length=512,
    strength=0.95,
    generator=torch.manual_seed(0)
)
output.values["images"][0].save("modular_flux_image.png")

[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.

[asomoza]

LGTM, it also fixed the issue I was having with the nodes!