app.py

import gradio as gr
from huggingface_hub import InferenceClient

# Switch to DialoGPT - no formatting token issues
client = InferenceClient("microsoft/DialoGPT-medium")

def respond(message, history):
    # Build conversation context
    conversation = ""
    if history:
        for user_msg, assistant_msg in history:
            conversation += f"User: {user_msg}\nAssistant: {assistant_msg}\n"
    
    conversation += f"User: {message}\nAssistant:"
    
    response = client.text_generation(
        conversation,
        max_new_tokens=100,
        temperature=0.1,
        return_full_text=False,
        stop_sequences=["\nUser:", "User:"]
    )
    
    yield response.strip()

chatbot = gr.ChatInterface(respond, type="messages")
chatbot.launch()