app.py

import torch
import gradio as gr
from config.model_config import ModelConfig
from src.data.tokenizer import CharacterTokenizer
from src.utils.helpers import generate, setup_logging

# Setup logging
logger = setup_logging()


def load_model():
    config = ModelConfig()
    device = "cuda" if torch.cuda.is_available() else "cpu"
    logger.info(f"Using device: {device}")

    # Load tokenizer
    with open(config.data_path) as f:
        text = f.read()
    tokenizer = CharacterTokenizer(text)

    # Load model
    try:
        model = torch.load(config.checkpoint_path, map_location=device)
        model.eval()
        return model, tokenizer, device
    except Exception as e:
        logger.error(f"Error loading model: {e}")
        raise


def generate_text(prompt, max_tokens=200, temperature=0.8):
    try:
        result = generate(model, tokenizer, prompt, max_tokens, device)
        return prompt + result
    except Exception as e:
        logger.error(f"Error during generation: {e}")
        return f"Error: {str(e)}"


# Load model globally
try:
    model, tokenizer, device = load_model()
    logger.info("Model loaded successfully")
except Exception as e:
    logger.error(f"Failed to load model: {e}")
    raise

# Create Gradio interface
demo = gr.Interface(
    fn=generate_text,
    inputs=[
        gr.Textbox(label="Enter your prompt", placeholder="Type your prompt here..."),
        gr.Slider(minimum=10, maximum=1000, value=200, step=10, label="Max Tokens"),
    ],
    outputs=gr.Textbox(label="Generated Text"),
    title="Shakespeare GPT",
    description="Enter a prompt and generate text using a custom GPT model",
    examples=[
        ["Hello, my name is", 200, 0.8],
        ["Once upon a time", 500, 0.8],
        ["The meaning of life is", 300, 0.8],
    ],
)

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