app.py

import gradio as gr
import numpy as np
import random
import torch
from PIL import Image
import os
from huggingface_hub import hf_hub_download
from pathlib import Path
import sys

# Add src directory to Python path
sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))

from src import model_loader
from src import pipeline
from src.config import Config, DeviceConfig
from transformers import CLIPTokenizer

# Create data directory if it doesn't exist
data_dir = Path("data")
data_dir.mkdir(exist_ok=True)

# Model configuration
MODEL_REPO = "stable-diffusion-v1-5/stable-diffusion-v1-5"
MODEL_FILENAME = "v1-5-pruned-emaonly.ckpt"
model_file = data_dir / MODEL_FILENAME

# Download model if it doesn't exist
if not model_file.exists():
    print(f"Downloading model from {MODEL_REPO}...")
    model_file = hf_hub_download(
        repo_id=MODEL_REPO,
        filename=MODEL_FILENAME,
        local_dir=data_dir,
        local_dir_use_symlinks=False
    )
    print("Model downloaded successfully!")

# Device configuration
device = "cuda" if torch.cuda.is_available() else "cpu"
print(f"Using device: {device}")

# Initialize configuration
config = Config(
    device=DeviceConfig(device=device),
    tokenizer=CLIPTokenizer.from_pretrained("openai/clip-vit-base-patch32")
)

# Load models with SE blocks enabled
config.models = model_loader.load_models(str(model_file), device, use_se=True)

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

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)
    
    # Update config with user settings
    config.seed = seed
    config.diffusion.cfg_scale = guidance_scale
    config.diffusion.n_inference_steps = num_inference_steps
    config.model.width = width
    config.model.height = height
    
    # Generate image
    output_image = pipeline.generate(
        prompt=prompt,
        uncond_prompt=negative_prompt,
        config=config
    )
    
    # Convert numpy array to PIL Image
    image = Image.fromarray(output_image)
    
    return image, seed

examples = [
    "A ultra sharp photorealtici painting of a futuristic cityscape at night with neon lights and flying cars",
    "A serene mountain landscape at sunset with snow-capped peaks and a clear lake reflection",
    "A detailed portrait of a cyberpunk character with glowing neon implants and holographic tattoos",
]

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(" # Custom Diffusion Model Text-to-Image Generator")

        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=42,
            )

            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,
                )

            with gr.Row():
                guidance_scale = gr.Slider(
                    label="Guidance scale",
                    minimum=0.0,
                    maximum=10.0,
                    step=0.1,
                    value=7.5,
                )

                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=[
            prompt,
            negative_prompt,
            seed,
            randomize_seed,
            width,
            height,
            guidance_scale,
            num_inference_steps,
        ],
        outputs=[result, seed],
    )

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