app.py

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

from diffusers import DiffusionPipeline
import torch
from huggingface_hub import login

import yaml

with open("config.yaml", "r") as f:
    config = yaml.safe_load(f)

token = config.get("huggingface_token")

# Login to Hugging Face Hub
login(token)

# Model details
device = "cuda" if torch.cuda.is_available() else "cpu"
torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32

# Load model
pipe = DiffusionPipeline.from_pretrained(
    "Grandediw/lora_model",
    torch_dtype=torch_dtype,
    use_auth_token=True  # Enables private model access
)
pipe = pipe.to(device)


# Constants
MAX_SEED = np.iinfo(np.int32).max
MAX_IMAGE_SIZE = 1024

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

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

    # Generate the image
    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

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

# Improved CSS for better styling
css = """
#interface-container {
    margin: 0 auto;
    max-width: 700px;
    padding: 10px;
    box-shadow: 0px 4px 10px rgba(0, 0, 0, 0.1);
    border-radius: 10px;
    background-color: #f9f9f9;
}
#header {
    text-align: center;
    font-size: 1.5em;
    margin-bottom: 20px;
    color: #333;
}
#advanced-settings {
    background-color: #f1f1f1;
    padding: 10px;
    border-radius: 8px;
}
"""

# Gradio interface
with gr.Blocks(css=css) as demo:
    with gr.Box(elem_id="interface-container"):
        gr.Markdown(
            """
            <div id="header">🖼️ Text-to-Image Generator</div>
            Generate high-quality images from your text prompts with the fine-tuned LoRA model.
            """
        )

        # Main input row
        with gr.Row():
            prompt = gr.Textbox(
                label="Prompt",
                placeholder="Describe the image you want to create...",
                lines=2,
            )
            run_button = gr.Button("Generate Image", variant="primary")

        # Output image display
        result = gr.Image(label="Generated Image").style(height="512px")

        # Advanced settings
        with gr.Accordion("Advanced Settings", open=False, elem_id="advanced-settings"):
            negative_prompt = gr.Textbox(
                label="Negative Prompt",
                placeholder="What to exclude from the image...",
            )
            randomize_seed = gr.Checkbox(label="Randomize Seed", value=True)
            seed = gr.Number(label="Seed", value=0, interactive=True)

            with gr.Row():
                width = gr.Slider(
                    label="Image Width",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=64,
                    value=512,
                )
                height = gr.Slider(
                    label="Image Height",
                    minimum=256,
                    maximum=MAX_IMAGE_SIZE,
                    step=64,
                    value=512,
                )

            with gr.Row():
                guidance_scale = gr.Slider(
                    label="Guidance Scale",
                    minimum=0.0,
                    maximum=20.0,
                    step=0.1,
                    value=7.5,
                )
                num_inference_steps = gr.Slider(
                    label="Steps",
                    minimum=10,
                    maximum=100,
                    step=5,
                    value=50,
                )

        # Examples
        gr.Examples(
            examples=examples,
            inputs=[prompt],
            outputs=[result],
            label="Try these prompts",
        )

    # Event handler
    run_button.click(
        fn=infer,
        inputs=[
            prompt,
            negative_prompt,
            seed,
            randomize_seed,
            width,
            height,
            guidance_scale,
            num_inference_steps,
        ],
        outputs=[result, seed],
    )

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