app.py

import gradio as gr
import numpy as np
import random
# import spaces #[uncomment to use ZeroGPU]
from diffusers import StableDiffusionPipeline
from peft import PeftModel, PeftConfig
import torch
from sympy.core.random import choice
from rembg import remove

device = "cuda" if torch.cuda.is_available() else "cpu"

MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 1024
# model_repo_id = "stabilityai/sdxl-turbo"  # Replace to the model you would like to use

# @spaces.GPU #[uncomment to use ZeroGPU]
def infer(
        model_id,
        lora,
        lora_scale,
        del_back,
        prompt,
        negative_prompt,
        seed,
        randomize_seed,
        width,
        height,
        guidance_scale,
        num_inference_steps,
        progress=gr.Progress(track_tqdm=True),
):
    if torch.cuda.is_available():
        torch_dtype = torch.float16
    else:
        torch_dtype = torch.float32

    pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch_dtype, use_safetensors=True)
    is_lora = False
    if model_id == "CompVis/stable-diffusion-v1-4" and lora == "pepe":
        lora_id = "seregasmirnov/pepe-lora"
        pipe.unet = PeftModel.from_pretrained(
            pipe.unet, 
            lora_id, 
            adapter_name="default"
          )
        is_lora = True
    pipe = pipe.to(device)

    if randomize_seed:
        seed = random.randint(0, MAX_SEED)

    generator = torch.Generator().manual_seed(seed)

    if is_lora:
        image = pipe(
            prompt=prompt,
            negative_prompt=negative_prompt,
            guidance_scale=guidance_scale,
            num_inference_steps=num_inference_steps,
            width=width,
            height=height,
            generator=generator,
            cross_attention_kwargs={"scale": lora_scale}
        ).images[0]
    else:
        image = pipe(
            prompt=prompt,
            negative_prompt=negative_prompt,
            guidance_scale=guidance_scale,
            num_inference_steps=num_inference_steps,
            width=width,
            height=height,
            generator=generator
        ).images[0]
        
    if del_back:
        image = remove(image)
        

    return image, seed


examples = [
    "sticker of a happy cat climbing a tree",
    "cute animal",
    "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
    "An astronaut riding a green horse",
    "A delicious ceviche cheesecake slice",
]
neg_examples = ["cat, dog",]

css = """
#col-container {
    margin: 0 auto;
    max-width: 640px;
}
"""

with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="col-container"):
        gr.Markdown(" # Text-to-Image Gradio Template")

        with gr.Row():
            model_id = gr.Dropdown(
                choices=["CompVis/stable-diffusion-v1-4", "stabilityai/sdxl-turbo", "stabilityai/stable-diffusion-xl-base-1.0"], info="Choose model")
        
            lora = gr.Dropdown(
                choices=["None", "pepe"], info="Choose lora", visible=True)

            lora_scale = gr.Slider(
                label="scale lora strength",
                minimum=0.0,
                maximum=2.0,
                step=0.1,
                value=1.0,
                visible=False,  
                info="setup lora strength"
            )

        def setup_lora(sel_model, sel_lora):
            if sel_model == "CompVis/stable-diffusion-v1-4":
                if sel_lora == "None":
                    return [gr.Dropdown(choices=["None", "pepe"], info="Choose lora", visible=True), gr.Slider(visible=False)]
                else:
                    return [gr.Dropdown(choices=["None", "pepe"], info="Choose lora", visible=True), gr.Slider(visible=True)]
            else:
                return [gr.Dropdown(choices=["None", "pepe"], info="Choose lora", visible=False), gr.Slider(visible=False)]
               
        model_id.change(
            fn=setup_lora,
            inputs=[model_id, lora],
            outputs=[lora, lora_scale])
        
        lora.change(
            fn=setup_lora,
            inputs=[model_id, lora],
            outputs=[lora, lora_scale])
        
        with gr.Row():
            del_back = gr.Checkbox(label="Delete background", value=False)

            prompt = gr.Text(
                label="Prompt",
                show_label=False,
                max_lines=1,
                placeholder="Enter your prompt",
                container=False,
                value="sticker of a happy cat climbing a tree",
            )

            run_button = gr.Button("Run", scale=0, variant="primary")

        result = gr.Image(label="Result", show_label=False)

        with gr.Accordion("Advanced Settings", open=False):
            negative_prompt = gr.Text(
                label="Negative prompt",
                max_lines=1,
                placeholder="Enter a negative prompt",
                visible=True,
                value="",
            )
            gr.Examples(examples=neg_examples, inputs=[negative_prompt])

            seed = gr.Slider(
                label="Seed",
                minimum=0,
                maximum=MAX_SEED,
                step=1,
                value=42,#0,
            )

            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)

            with gr.Row():
                width = gr.Slider(
                    label="Width",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=512,  
                )

                height = gr.Slider(
                    label="Height",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=512,  
                )

                guidance_scale = gr.Slider(
                    label="Guidance scale",
                    minimum=0.0,
                    maximum=10.0,
                    step=0.1,
                    value=4.0,
                )

                num_inference_steps = gr.Slider(
                    label="Number of inference steps",
                    minimum=1,
                    maximum=50,
                    step=1,
                    value=50,  
                )

        gr.Examples(examples=examples, inputs=[prompt])

    gr.on(
        triggers=[run_button.click, prompt.submit],
        fn=infer,
        inputs=[
            model_id,
            lora,
            lora_scale,
            del_back,
            prompt,
            negative_prompt,
            seed,
            randomize_seed,
            width,
            height,
            guidance_scale,
            num_inference_steps,
        ],
        outputs=[result, seed],
    )

if __name__ == "__main__":
    demo.launch()