app.py

import asyncio
from transformers import AutoTokenizer, AutoModelForCausalLM
from datasets import load_dataset

# -------------------------------
# Model loading
# -------------------------------
MODEL_NAME = "LiquidAI/LFM2-2.6B"

print("Loading model...")
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
model = AutoModelForCausalLM.from_pretrained(MODEL_NAME)
print("Model loaded.")

# -------------------------------
# Fetch prompts from Hugging Face
# -------------------------------
async def fetch_prompts():
    print("Fetching prompts from Hugging Face dataset...")
    dataset = load_dataset("fka/awesome-chatgpt-prompts", split="train")
    all_prompts = dataset['prompt']  # List of 200+ prompts
    print(f"Total prompts available: {len(all_prompts)}")
    return all_prompts

# -------------------------------
# Main async function
# -------------------------------
async def main():
    all_prompts = await fetch_prompts()

    # Use first 20 for fast startup
    fast_prompts = all_prompts[:20]
    print("Using first 20 prompts for fast startup...")
    for i, p in enumerate(fast_prompts, 1):
        print(f"[Prompt {i}] {p}")

    # Load remaining prompts asynchronously
    remaining_prompts = all_prompts[20:]
    print(f"Loading remaining {len(remaining_prompts)} prompts asynchronously...")

    # Example: You could process remaining prompts in background
    # For demonstration, just print count after "async load"
    await asyncio.sleep(1)
    print("Remaining prompts loaded.")

# -------------------------------
# Entry point
# -------------------------------
if __name__ == "__main__":
    asyncio.run(main())