app.py

import os
import gradio as gr
import torch
import numpy as np
from PIL import Image

from diffusers import QwenImageEditPlusPipeline

model_path = "ovedrive/Qwen-Image-Edit-2509-4bit"
pipeline = QwenImageEditPlusPipeline.from_pretrained(model_path, torch_dtype=torch.bfloat16)
print("pipeline loaded") # not true but whatever. do not move to cuda

pipeline.set_progress_bar_config(disable=None)
#pipeline.enable_model_cpu_offload() #if you have enough VRAM replace this line with `pipeline.to("cuda")` which is 20GB VRAM
pipeline.to("cuda")
image = None
prompt = None

def generate(image: Image.Image, prompt: str):
    inputs = {
        "image": image,
        "prompt": prompt,
        "generator": torch.manual_seed(0),
        "true_cfg_scale": 4.0,
        "negative_prompt": " ",
        "num_inference_steps": 20, # even 10 steps should be enough in many cases
    }
    
    result = None

    with torch.inference_mode():
        output = pipeline(**inputs)
        output_image = output.images[0]
        output_image.save("output_image_edit_plus.png")
        print("image saved at", os.path.abspath("output_image_edit_plus.png"))
    
    return os.path.abspath("output_image_edit_plus.png")


with gr.Blocks(title="teste") as demo:
        image = gr.Image(type="pil", label="Imagem de entrada")
        prompt = gr.Textbox(label="Edit prompt", placeholder="Descreva a edição desejada...")
        run = gr.Button("Gerar")
        out = gr.Image(label="Saída")
        run.click(fn=generate, inputs=[image, prompt], outputs=[out])
    

if __name__ == "__main__":
    demo.queue().launch(
        server_name=os.getenv("GRADIO_SERVER_NAME", "0.0.0.0"),
        server_port=int(os.getenv("GRADIO_SERVER_PORT", "7861")),
        show_error=True,
    )