Flux.2-klein-4b lora support

[christopher5106]

Describe the bug

Not sure about the origin of this lora (fal trainer?)

Reproduction

import torch
from diffusers import Flux2KleinPipeline

pipe = Flux2KleinPipeline.from_pretrained("black-forest-labs/FLUX.2-klein-4b", torch_dtype=torch.bfloat16)
pipe.load_lora_weights(
    "fal/flux-2-klein-4b-spritesheet-lora", weight_name="flux-spritesheet-lora.safetensors",
)

Logs

No LoRA keys associated to Flux2Transformer2DModel found with the prefix='transformer'. This is safe to ignore if LoRA state dict didn't originally have any Flux2Transformer2DModel related params. You can also try specifying `prefix=None` to resolve the warning. Otherwise, open an issue if you think it's unexpected: https://github.com/huggingface/diffusers/issues/new

System Info

Diffusers version from today 6875490

Who can help?

@sayakpaul

[ranag786tech-lang]

``# 1. Ensure your library is up-to-date

pip install -U diffusers

2. Use this updated load call:

pipe.load_lora_weights(
"fai/flux-2-klein-4b-spritesheet-lora",
weight_name="flux-spritesheet-lora.safetensors",
adapter_name="spritesheet"
)

3. Explicitly set the adapter to activate it

pipe.set_adapters(["spritesheet"], adapter_weights=[0.9])

Update your loading code to include the weight_name and an adapter_name This forces the mapping and resolves the "No LoRA keys associated" warning

[christopher5106]

Flux2 klein currently not on diffusers' pypi, so your answer cannot be correct.

[christopher5106]

Others having same issue with loras from Fal trainer:

https://huggingface.co/fal/flux-2-klein-4B-object-remove-lora/discussions/3

[asomoza]

yeah, that lora uses the peft key naming instead of the diffusers one. I'll open a PR.

BTW very cool lora.