app.py

import gradio as gr
from huggingface_hub import InferenceClient
import os

token = os.getenv("huggingface_token")
client = InferenceClient("HuggingFaceH4/zephyr-7b-beta", token=token)

def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages = [{"role": "system", "content": system_message}]
    for user, assistant in history:
        if user:
            messages.append({"role": "user", "content": user})
        if assistant:
            messages.append({"role": "assistant", "content": assistant})
    messages.append({"role": "user", "content": message})

    response = ""
    for message in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token_piece = message.choices[0].delta.content
        response += token_piece
        yield response

# Make a version of `respond` that's API-callable
def chat_api(message, max_tokens, temperature, top_p):
    # Create a minimal prompt with default system message and no chat history
    messages = [{"role": "system", "content": DEFAULT_SYSTEM_MESSAGE}]
    messages.append({"role": "user", "content": message})

    output = ""
    for chunk in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        output += chunk.choices[0].delta.content
    return output

DEFAULT_SYSTEM_MESSAGE = (
    "I am Bilguun, a third year software engineering student at McGill University and the founder of VentureCircle. "
    "VentureCircle is based in Montreal. "
    "I created this chatbot, LaunchPad, to help aspiring entrepreneurs achieve their goals. "
    "Your name is LaunchPad. "
    "LaunchPad is a friendly AI assistant designed to provide useful and supportive guidance for starting and growing a startup."
)

# Main demo UI
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value=DEFAULT_SYSTEM_MESSAGE, label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
    ],
)

# Launch with /chat endpoint
demo.launch(share=True, show_error=True).api(fn=chat_api, inputs=["text", "slider", "slider", "slider"], api_name="/chat")