app.py

import gradio as gr
from openai import OpenAI

client = OpenAI(
    base_url="https://integrate.api.nvidia.com/v1",
    api_key="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"
)

system_prompt = {"role": "system", "content": "You are a helpful assistant to answer user queries."}

def get_text_response(user_message, history=None):
    if history is None:
        history = []

    # Prepare messages for OpenAI API in the correct format
    messages = [system_prompt]
    for user_msg, assistant_msg in history:
        messages.append({"role": "user", "content": user_msg})
        messages.append({"role": "assistant", "content": assistant_msg})
    messages.append({"role": "user", "content": user_message})

    response = ""
    completion = client.chat.completions.create(
        model="nvidia/llama-3.1-nemotron-70b-instruct",
        messages=messages,
        temperature=0.5,
        top_p=1,
        max_tokens=100,
        stream=True
    )

    for chunk in completion:
        delta = chunk.choices[0].delta
        if delta and delta.content:
            response += delta.content

    history.append((user_message, response))
    return history, history

demo = gr.ChatInterface(
    fn=get_text_response,
    title="🧠 Nemotron 70B Assistant",
    theme="soft",
    examples=["How are you doing?", "What are your interests?", "Which places do you like to visit?"]
)

if __name__ == "__main__":
    demo.queue().launch(share=True, debug=True)