app.py

import gradio as gr
import os

from diffusers import DiffusionPipeline, StableDiffusionXLImg2ImgPipeline
import torch
import random
import uuid

#token = os.getenv("token")
#model = gr.load("models/Rojban/dreambooth4", hf_token=token)
prj_path = "dreambooth4"

model = "stabilityai/stable-diffusion-xl-base-1.0"

pipe = DiffusionPipeline.from_pretrained(
    model,
    torch_dtype=torch.float16,
)
pipe.to("cuda")
pipe.load_lora_weights(prj_path, weight_name="pytorch_lora_weights.safetensors")

refiner = StableDiffusionXLImg2ImgPipeline.from_pretrained(
    "stabilityai/stable-diffusion-xl-refiner-1.0",
   torch_dtype=torch.float16,
)
refiner.to("cuda")


def generate_image(prompt, seed=None):
    if seed is None:
        seed = 253

    seed = int(seed)
    generator = torch.Generator("cuda").manual_seed(seed)
    image = pipe(prompt=prompt, generator=generator, num_inference_steps=25).images[0]
    image = refiner(prompt=prompt, generator=generator, image=image).images[0]
    name = f"{seed}_{str(uuid.uuid4())}.png"
    save_path = f"images/{name}"
    image.save(save_path)
    return save_path


# Create the Gradio interface
interface = gr.Interface(
    fn=generate_image,
    inputs=[gr.Textbox(label="Prompt"), gr.Number(label="Seed")],
    outputs=gr.Image(type="filepath"),
    title="Custom Stable Diffusion Model",
    description="Generate images using a custom Stable Diffusion model.",
)

# Launch the app
if __name__ == "__main__":
    interface.launch()