app.py

import gradio as gr
from huggingface_hub import InferenceClient

# Initialize the client with your model
client = InferenceClient("Arnic/gemma2-2b-it-Pubmed20k-TPU")

# Define response function
def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    system_message = (
        "You are a good listener. You advise relaxation exercises, suggest avoiding negative thoughts, "
        "and guide through steps to manage stress. Let's discuss what's on your mind, "
        "or ask me for a quick relaxation exercise."
    )

    # Format history and system message as prompt text
    chat_history = ""
    for user_msg, bot_reply in history:
        if user_msg:
            chat_history += f"User: {user_msg}\n"
        if bot_reply:
            chat_history += f"Assistant: {bot_reply}\n"
    
    prompt = f"{system_message}\n\n{chat_history}User: {message}\nAssistant:"

    # Generate response using the InferenceClient text generation method
    response = client.text_generation(
        prompt=prompt,
        max_new_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p
    )

    # Extract and yield the text response
    generated_text = response["generated_text"].replace(prompt, "").strip()
    yield generated_text

# Set up Gradio interface
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", 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)",
        ),
    ],
)

if __name__ == "__main__":
    demo.launch()