app.py

import spaces
from diffusers import ControlNetModel, StableDiffusionXLControlNetPipeline, AutoencoderKL, EulerAncestralDiscreteScheduler
from PIL import Image
import torch
import numpy as np
import cv2
import gradio as gr
from torchvision import transforms 

controlnet = ControlNetModel.from_pretrained(
    "briaai/BRIA-2.2-ControlNet-Recoloring",
    torch_dtype=torch.float16
).to('cuda')

pipe = StableDiffusionXLControlNetPipeline.from_pretrained(
    "briaai/BRIA-2.2",
    controlnet=controlnet,
    torch_dtype=torch.float16,
    low_cpu_mem_usage=True,
    offload_state_dict=True,
).to('cuda').to(torch.float16)

pipe.scheduler = EulerAncestralDiscreteScheduler(
    beta_start=0.00085,
    beta_end=0.012,
    beta_schedule="scaled_linear",
    num_train_timesteps=1000,
    steps_offset=1
)
pipe.force_zeros_for_empty_prompt = False

def resize_image(image):
    image = image.convert('RGB')
    current_size = image.size
    if current_size[0] > current_size[1]:
        center_cropped_image = transforms.functional.center_crop(image, (current_size[1], current_size[1]))
    else:
        center_cropped_image = transforms.functional.center_crop(image, (current_size[0], current_size[0]))
    resized_image = transforms.functional.resize(center_cropped_image, (1024, 1024))
    return resized_image

@spaces.GPU
def generate_(prompt, negative_prompt, grayscale_image, num_steps, controlnet_conditioning_scale, seed):
    generator = torch.Generator("cuda").manual_seed(seed)    
    images = pipe(
        prompt, 
        negative_prompt=negative_prompt, 
        image=grayscale_image, 
        num_inference_steps=num_steps, 
        controlnet_conditioning_scale=float(controlnet_conditioning_scale),
        generator=generator,
    ).images
    return images

@spaces.GPU
def process(input_image, prompt, negative_prompt, num_steps, controlnet_conditioning_scale, seed):
    input_image = resize_image(input_image)
    grayscale_image = input_image.convert('L').convert('RGB')
    images = generate_(prompt, negative_prompt, grayscale_image, num_steps, controlnet_conditioning_scale, seed)
    return grayscale_image, images[0]

block = gr.Blocks()

with block:
    gr.Markdown("## BRIA 2.2 ControlNet Recoloring")
    gr.HTML('''
      <p style="margin-bottom: 10px; font-size: 94%">
        This is a demo for ControlNet Recoloring that uses
        <a href="https://huggingface.co/briaai/BRIA-2.2" target="_blank">BRIA 2.2 text-to-image model</a> as backbone. 
        Trained on licensed data, BRIA 2.2 provides full legal liability coverage for copyright and privacy infringement.
      </p>
    ''')
    with gr.Row():
        with gr.Column():
            input_image = gr.Image(sources=None, type="pil")
            prompt = gr.Textbox(label="Prompt")
            negative_prompt = gr.Textbox(label="Negative prompt", value="Logo,Watermark,Text,Ugly,Morbid,Extra fingers,Poorly drawn hands,Mutation,Blurry,Extra limbs,Gross proportions,Missing arms,Mutated hands,Long neck,Duplicate,Mutilated,Mutilated hands,Poorly drawn face,Deformed,Bad anatomy,Cloned face,Malformed limbs,Missing legs,Too many fingers")
            num_steps = gr.Slider(label="Number of steps", minimum=25, maximum=100, value=50, step=1)
            controlnet_conditioning_scale = gr.Slider(label="ControlNet conditioning scale", minimum=0.1, maximum=2.0, value=1.0, step=0.05)
            seed = gr.Slider(label="Seed", minimum=0, maximum=2147483647, step=1, randomize=True)
            run_button = gr.Button(value="Run")
        with gr.Column():
            preview = gr.Image(label="Control (Grayscale)", type="pil")
            result = gr.Image(label="Output Image", type="pil")
    ips = [input_image, prompt, negative_prompt, num_steps, controlnet_conditioning_scale, seed]
    run_button.click(fn=process, inputs=ips, outputs=[preview, result])

block.launch(debug=True)