app.py

import gradio as gr
from transformers import pipeline


generator = pipeline("text-generation", model="microsoft/phi-4-reasoning-plus", device_map="auto")

def chat_with_phi(prompt):
    response = generator(prompt, max_new_tokens=200, do_sample=False)[0]["generated_text"]
    return response

demo = gr.Interface(fn=chat_with_phi, inputs="text", outputs="text")
demo.launch()