app.py

import gradio as gr
import os
import sys
import subprocess
import numpy as np
from PIL import Image
import cv2

import torch
from diffusers import StableDiffusion3Pipeline
from diffusers.models.controlnet_sd3 import ControlNetSD3Model
from diffusers.utils.torch_utils import randn_tensor
from diffusers.utils import load_image

# Clone the specific branch
subprocess.run(["git", "clone", "-b", "sd3_control", "https://github.com/instantX-research/diffusers_sd3_control.git"])

# Change directory to the cloned repository and install it
os.chdir('diffusers_sd3_control')
subprocess.run(["pip", "install", "-e", "."])

# Add the path to the examples directory
sys.path.append(os.path.abspath('./examples/community'))

# Import the required pipeline
from pipeline_stable_diffusion_3_controlnet import StableDiffusion3CommonPipeline



# load pipeline
base_model = 'stabilityai/stable-diffusion-3-medium-diffusers'
pipe = StableDiffusion3CommonPipeline.from_pretrained(
    base_model, 
    controlnet_list=['InstantX/SD3-Controlnet-Canny'],
)
pipe.to('cuda:0', torch.float16)

def resize_image(input_path, output_path, target_height):
    # Open the input image
    img = Image.open(input_path)

    # Calculate the aspect ratio of the original image
    original_width, original_height = img.size
    original_aspect_ratio = original_width / original_height

    # Calculate the new width while maintaining the aspect ratio and the target height
    new_width = int(target_height * original_aspect_ratio)

    # Resize the image while maintaining the aspect ratio and fixing the height
    img = img.resize((new_width, target_height), Image.LANCZOS)

    # Save the resized image
    img.save(output_path)

    return output_path, new_width, target_height

def infer(image_in, prompt, inference_steps, guidance_scale, control_weight):
    
    n_prompt = 'NSFW, nude, naked, porn, ugly'

    image_in, w, h = resize_image(image_in, "resized_input.jpg", 1024)
    
    # Canny preprocessing
    image_to_canny = load_image(image_in)
    image_to_canny = np.array(image_to_canny)
    image_to_canny = cv2.Canny(image_to_canny, 100, 200)
    image_to_canny = image_to_canny[:, :, None]
    image_to_canny = np.concatenate([image_to_canny, image_to_canny, image_to_canny], axis=2)
    image_to_canny = Image.fromarray(image_to_canny)
    
    # controlnet config
    controlnet_conditioning = [
        dict(
            control_index=0,
            control_image=image_to_canny,
            control_weight=control_weight,
            control_pooled_projections='zeros'
        )
    ]
    
    # infer
    image = pipe(
        prompt=prompt,
        negative_prompt=n_prompt,
        controlnet_conditioning=controlnet_conditioning,
        num_inference_steps=inference_steps,
        guidance_scale=guidance_scale,
        width=w,
        height=h
    ).images[0]

    return image, image_to_canny

css="""
#col-container{
    margin: 0 auto;
    max-width: 1080px;
}
"""
with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="col-container"):
        gr.Markdown("""
        # SD3 ControlNet
        """)
        with gr.Row():
            with gr.Column():
                image_in = gr.Image(label="Image reference", sources=["upload"], type="filepath")
                prompt = gr.Textbox(label="Prompt")
                with gr.Accordion("Advanced settings", open=False):
                    with gr.Column():
                        with gr.Row():
                            inference_steps = gr.Slider(label="Inference steps", minimum=1, maximum=50, step=1, value=25)
                            guidance_scale = gr.Slider(label="Guidance scale", minimum=1.0, maximum=10.0, step=0.1, value=7.0)
                        control_weight = gr.Slider(label="Control Weight", minimum=0.0, maximum=1.0, step=0.01, value=0.7)
                    
                submit_btn = gr.Button("Submit")
            with gr.Column():
                result = gr.Image(label="Result")
                canny_used = gr.Image(label="Preprocessed Canny")
    
    submit_btn.click(
        fn = infer,
        inputs = [image_in, prompt, inference_steps, guidance_scale, control_weight],
        outputs = [result, canny_used],
        show_api=False
    )
demo.queue().launch()