app.py

#refer llama recipes for more info https://github.com/huggingface/huggingface-llama-recipes/blob/main/inference-api.ipynb
#huggingface-llama-recipes : https://github.com/huggingface/huggingface-llama-recipes/tree/main
import gradio as gr
from openai import OpenAI
import os

css = '''
.gradio-container{max-width: 1000px !important}
h1{text-align:center}
footer {
    visibility: hidden
}
'''

ACCESS_TOKEN = os.getenv("HF_TOKEN")

client = OpenAI(
    base_url="https://api-inference.huggingface.co/v1/",
    api_key=ACCESS_TOKEN,
)

def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages = [{"role": "system", "content": system_message}]

    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

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

    response = ""
    
    for message in  client.chat.completions.create(
        model="meta-llama/Meta-Llama-3.1-8B-Instruct",
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
        messages=messages,
    ):
        token = message.choices[0].delta.content

        response += token
        yield response

demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="", label="System message", visible=False),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens", visible=False),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature", visible=False),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-P",
            visible=False,
        ),
        
    ],
    css=css,
    theme="allenai/gradio-theme",
)
if __name__ == "__main__":
    demo.launch()