app.py

import gradio as gr
import numpy as np
import random
import torch
from diffusers import DiffusionPipeline, EDMEulerScheduler

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

# 🎯 Model ID and config
model_repo_id = "stabilityai/sdxl-turbo"
MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 1024

# 🔁 Load model with EDM + VPred scheduler
pipe = DiffusionPipeline.from_pretrained(
    model_repo_id,
    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
    variant="fp16" if torch.cuda.is_available() else None,
)

# 🔄 Replace scheduler with EDM + V-prediction
pipe.scheduler = EDMEulerScheduler.from_config(pipe.scheduler.config)

# 🧠 Enable optimizations if GPU
if device == "cuda":
    try:
        pipe.enable_xformers_memory_efficient_attention()
    except Exception as e:
        print("⚠️ xFormers not available, using attention slicing.")
        pipe.enable_attention_slicing()

    pipe.enable_model_cpu_offload()
    pipe.enable_vae_tiling()

pipe = pipe.to(device)

# 🚀 Inference function
def infer(
    prompt,
    negative_prompt,
    seed,
    randomize_seed,
    width,
    height,
    guidance_scale,
    num_inference_steps,
    progress=gr.Progress(track_tqdm=True),
):
    if randomize_seed:
        seed = random.randint(0, MAX_SEED)

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

    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]

    return image, seed

# 🧪 Prompt examples
examples = [
    "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
    "An astronaut riding a green horse",
    "A delicious ceviche cheesecake slice",
]

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

# 🧱 Gradio Interface
with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="col-container"):
        gr.Markdown(" # Text-to-Image Gradio (EDM + VPred)")

        with gr.Row():
            prompt = gr.Text(
                label="Prompt",
                show_label=False,
                max_lines=1,
                placeholder="Enter your prompt",
                container=False,
            )

            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=False,
            )

            seed = gr.Slider(
                label="Seed",
                minimum=0,
                maximum=MAX_SEED,
                step=1,
                value=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=1024,
                )

                height = gr.Slider(
                    label="Height",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=32,
                    value=1024,
                )

            with gr.Row():
                guidance_scale = gr.Slider(
                    label="Guidance scale",
                    minimum=0.0,
                    maximum=10.0,
                    step=0.1,
                    value=2.5,  # Optimal for SDXL-Turbo
                )

                num_inference_steps = gr.Slider(
                    label="Number of inference steps",
                    minimum=1,
                    maximum=20,
                    step=1,
                    value=4,  # Low default for EDM
                )

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

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

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