app.py

import spaces
import gradio as gr
from huggingface_hub import InferenceClient

from vptq.app_utils import get_chat_loop_generator

# Update model list with annotations
model_list_with_annotations = {
    # "VPTQ-community/Meta-Llama-3.1-70B-Instruct-v8-k65536-65536-woft": "Llama 3.1 70B @ 4bit",
    # "VPTQ-community/Meta-Llama-3.1-70B-Instruct-v8-k65536-256-woft": "Llama 3.1 70B @ 3bit",
    # "VPTQ-community/Meta-Llama-3.1-70B-Instruct-v16-k65536-65536-woft": "Llama 3.1 70B @ 2bit",
    # "VPTQ-community/Qwen2.5-72B-Instruct-v8-k65536-65536-woft": "Qwen2.5 72B @ 4 bits",
    # "VPTQ-community/Qwen2.5-72B-Instruct-v8-k65536-256-woft": "Qwen2.5 72B @ 3 bits",
    # "VPTQ-community/Qwen2.5-72B-Instruct-v16-k65536-65536-woft": "Qwen2.5 72B @ 3 bits",
    # "VPTQ-community/Qwen2.5-32B-Instruct-v8-k65536-65536-woft": "Qwen2.5 32B @ 4 bits",
    "VPTQ-community/Qwen2.5-32B-Instruct-v8-k65536-256-woft": "Qwen2.5 32B @ 3 bits",
    "VPTQ-community/Qwen2.5-32B-Instruct-v16-k65536-0-woft": "Qwen2.5 32B @ 2 bits"
}

# Create a list of choices with annotations for the dropdown
model_list_with_annotations_display = [f"{key} ({value})" for key, value in model_list_with_annotations.items()]

model_keys = list(model_list_with_annotations.keys())
current_model_g = model_keys[0]
chat_completion = get_chat_loop_generator(current_model_g)

@spaces.GPU
def update_title_and_chatmodel(model):
    model = str(model)
    global chat_completion
    global current_model_g
    if model != current_model_g:
        current_model_g = model
        chat_completion = get_chat_loop_generator(current_model_g)
    return model


@spaces.GPU
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 chat_completion(
            messages,
            max_tokens=max_tokens,
            stream=True,
            temperature=temperature,
            top_p=top_p,
    ):
        token = message

        response += token
        yield response


css = """
h1 {
  text-align: center;
  display: block;
}
"""

chatbot = gr.Chatbot(label="Gradio ChatInterface")
with gr.Blocks() as demo:
    with gr.Column(scale=1):
        title_output = gr.Markdown("Please select a model to run")
        chat_demo = gr.ChatInterface(
            respond,
            additional_inputs_accordion=gr.Accordion(
                label="⚙️ Parameters", open=False, render=False
            ),
            fill_height=False,
            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)"
                ),
            ],
        )
        model_select = gr.Dropdown(
            choices=model_list_with_annotations_display,
            label="Models",
            value=model_list_with_annotations_display[0],
            info="Model & Estimated Quantized Bitwidth"
        )
        model_select.change(update_title_and_chatmodel, inputs=[model_select], outputs=title_output)

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