app.py

from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
import gradio as gr

model_name = "microsoft/Phi-4-mini-instruct"

# Load model & tokenizer with optimizations
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")

# Create a pipeline for text generation (faster inference)
chatbot = pipeline("text-generation", model=model, tokenizer=tokenizer, max_new_tokens=200)

def chatbot_response(user_input):
    response = chatbot(user_input)[0]["generated_text"]
    return response

# Gradio UI
iface = gr.Interface(
    fn=chatbot_response,
    inputs="text",
    outputs="text",
    title="Ethical AI Chatbot",
    description="A chatbot for ethical AI guidance."
)

iface.launch()