app.py

import gradio as gr
from huggingface_hub import InferenceClient, login
from json import loads

import random

from os import getenv 

# Login to Hugging Face
login(getenv("Token")) 


client = InferenceClient(
    provider="nebius",
    model="meta-llama/Llama-3.2-3B-Instruct",
)#"meta-llama/Llama-3.2-3B-Instruct")

#Gradio's history is never used
def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
    actual_history
):
    #parse the actual history, which we received from the frontend, to JSON.
    #Overwrite Gradio history
    print('THE ACTUAL HISTORY', actual_history)

    history=loads(actual_history) if actual_history else []

    messages =[{"role": "system", "content": system_message}, *history] 


    messages.append({"role": "user", "content": message})

    
    print('the messages',messages)

    response = ""

    
    seed = random.randint(1, 10000)

    for message in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
        seed=seed
    ):
        print('the response',message)
        token = message.choices[0].delta.content

        response += token
        yield response

 
demo = gr.ChatInterface(
    respond,
    type="messages",
    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)",
        ), 
        gr.Textbox(value="", label="Actual history"),
    ],
)


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