app.py

import gradio as gr
from transformers import pipeline

# Load open-access model
pipe = pipeline("text-generation", model="tiiuae/falcon-7b-instruct", device_map="auto")

def ask(question):
    prompt = f"User: {question}\nAssistant:"
    response = pipe(prompt, max_new_tokens=200, do_sample=True, temperature=0.7)[0]['generated_text']
    return response.split("Assistant:")[-1].strip()

demo = gr.Interface(fn=ask, inputs="text", outputs="text",
                    title="🧠 Ask This LLM!",
                    description="Ask about any topic.")

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