Skip to content

make flux ready for mellon #12419

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 6 commits into from
Oct 6, 2025
Merged

make flux ready for mellon #12419

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
9dbed17
make flux ready for mellon
sayakpaul Oct 2, 2025
c67f59f
Merge branch 'main' into fix-flux-mellon
sayakpaul Oct 4, 2025
3efbdb2
Merge branch 'main' into fix-flux-mellon
sayakpaul Oct 5, 2025
7941216
Merge branch 'main' into fix-flux-mellon
sayakpaul Oct 6, 2025
a1edf8e
up
sayakpaul Oct 6, 2025
9861f5b
Apply suggestions from code review
sayakpaul Oct 6, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions src/diffusers/modular_pipelines/flux/encoders.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -181,6 +181,7 @@ def inputs(self) -> List[InputParam]:
return [
InputParam("prompt"),
InputParam("prompt_2"),
InputParam("max_sequence_length", type_hint=int, default=512, required=False),
Copy link
Member Author

Choose a reason for hiding this comment

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

Took the liberty of supporting this input from the user so that Schnell can also work. In another PR, I will harmonize the steps in Flux Modular that include repetition along the batch size dimension (similar to Qwen).

Cc: @yiyixuxu

InputParam("joint_attention_kwargs"),
]

Expand Down Expand Up @@ -404,6 +405,7 @@ def __call__(self, components: FluxModularPipeline, state: PipelineState) -> Pip
pooled_prompt_embeds=None,
device=block_state.device,
num_images_per_prompt=1, # TODO: hardcoded for now.
max_sequence_length=block_state.max_sequence_length,
lora_scale=block_state.text_encoder_lora_scale,
)

Expand Down
6 changes: 3 additions & 3 deletions src/diffusers/modular_pipelines/flux/modular_blocks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,9 +84,9 @@ def description(self):

# before_denoise: all task (text2img, img2img)
class FluxAutoBeforeDenoiseStep(AutoPipelineBlocks):
block_classes = [FluxBeforeDenoiseStep, FluxImg2ImgBeforeDenoiseStep]
block_names = ["text2image", "img2img"]
block_trigger_inputs = [None, "image_latents"]
block_classes = [FluxImg2ImgBeforeDenoiseStep, FluxBeforeDenoiseStep]
block_names = ["img2img", "text2image"]
block_trigger_inputs = ["image_latents", None]

@property
def description(self):
Expand Down
Loading