LibreFLUX-ControlNet Inference (WIP)

ControlNet trained on top of LibreFLUX

Uses Attention Masking
Inference runs with CFG
Trained on 165K segmented images from Meta's SA1B Dataset
Trained using https://github.com/NeuralVFX/LibreFLUX-ControlNet

Disclaimer, This is pieced together by modifying and borrowing from these repos:

Compatibility

Use these specific library versions:

pip install -U diffusers==0.32.0
pip install -U "transformers @ git+https://github.com/huggingface/transformers@e15687fffe5c9d20598a19aeab721ae0a7580f8a"

Load Pipeline

import torch
from diffusers import DiffusionPipeline

model_id = "neuralvfx/LibreFlux-ControlNet"  
device = "cuda" if torch.cuda.is_available() else "cpu"

dtype  = torch.bfloat16 if device == "cuda" else torch.float32

pipe = DiffusionPipeline.from_pretrained(
    model_id,
    custom_pipeline=model_id,
    trust_remote_code=True,   
    torch_dtype=dtype,
    safety_checker=None        
).to(device)

Inference

from PIL import Image
# Load Control Image
cond = Image.open("examples/libe_flux_control.png")
cond = cond.resize((1024, 1024))

# Convert PIL image to tensor and move to device with correct dtype
cond_tensor = ToTensor()(cond)[:3,:,:].to(pipe.device, dtype=pipe.dtype).unsqueeze(0)

out = pipe(
  prompt="many pieces of drift wood spelling libre flux sitting casting shadow on the lumpy sandy beach with foot prints all over it",
            negative_prompt="blurry",
            control_image=cond_tensor,  # Use the tensor here
            num_inference_steps=75,
            guidance_scale=4.0,
            height =1024,
            width=1024,
            controlnet_conditioning_scale=1.0,
            num_images_per_prompt=1,
            control_mode=None,
            generator= torch.Generator().manual_seed(32),
            return_dict=True,
        )
out.images[0]