app.py

import os
import requests
import json

import gradio as gr
from transformers import AutoTokenizer


DESCRIPTION = """

"""

LICENSE = """

"""

DEFAULT_SYSTEM_PROMPT = ""

API_URL = os.environ.get("API_URL")
TOKEN = os.environ.get("TOKEN")

HEADERS = {
    "accept": "application/json",
    "Authorization": f"Bearer {TOKEN}", "Content-Type": "application/json",
}

MODEL_NAME="breeze-7b-instruct-v01"
TEMPERATURE=1
MAX_TOKENS=16
TOP_P=0
PRESENCE_PENALTY=0
FREQUENCY_PENALTY=0


eos_token = "</s>"
MAX_MAX_NEW_TOKENS = 4096
DEFAULT_MAX_NEW_TOKENS = 1536

max_prompt_length = 8192 - MAX_MAX_NEW_TOKENS - 10

model_name = "MediaTek-Research/Breeze-7B-Instruct-v0.1"
tokenizer = AutoTokenizer.from_pretrained(model_name)

with gr.Blocks() as demo:
    gr.Markdown(DESCRIPTION)

    chatbot = gr.Chatbot()
    with gr.Row():
        msg = gr.Textbox(
            container=False,
            show_label=False,
            placeholder='Type a message...',
            scale=10,
        )
        submit_button = gr.Button('Submit',
                                  variant='primary',
                                  scale=1,
                                  min_width=0)

    with gr.Row():
        retry_button = gr.Button('🔄  Retry', variant='secondary')
        undo_button = gr.Button('↩️ Undo', variant='secondary')
        clear = gr.Button('🗑️  Clear', variant='secondary')

    saved_input = gr.State()

    with gr.Accordion(label='Advanced options', open=False):
        system_prompt = gr.Textbox(label='System prompt',
                                   value=DEFAULT_SYSTEM_PROMPT,
                                   lines=6)
        max_new_tokens = gr.Slider(
            label='Max new tokens',
            minimum=1,
            maximum=MAX_MAX_NEW_TOKENS,
            step=1,
            value=DEFAULT_MAX_NEW_TOKENS,
        )
        temperature = gr.Slider(
            label='Temperature',
            minimum=0.1,
            maximum=1.0,
            step=0.1,
            value=0.3,
        )
        top_p = gr.Slider(
            label='Top-p (nucleus sampling)',
            minimum=0.05,
            maximum=1.0,
            step=0.05,
            value=0.95,
        )
        top_k = gr.Slider(
            label='Top-k',
            minimum=1,
            maximum=1000,
            step=1,
            value=50,
        )

    def user(user_message, history):
        return "", history + [[user_message, None]]


    def bot(history, max_new_tokens, temperature, top_p, top_k, system_prompt):
        chat_data = []
        for user_msg, assistant_msg in history:
            if user_msg is not None:
                chat_data.append({"role": "user", "content": user_msg})
            if assistant_msg is not None:
                chat_data.append({"role": "assistant", "content": assistant_msg})

        message = tokenizer.apply_chat_template(chat_data, tokenize=False)
        message = message[3:]  # remove SOT token

        print(message)

        data = {
            "model": MODEL_NAME,
            "messages": str(message),
            "temperature": TEMPERATURE,
            "n": 1,
            "max_tokens": MAX_TOKENS,
            "stop": "",
            "top_p": TOP_P,
            "logprobs": 0,
            "echo": False,
            "presence_penalty": PRESENCE_PENALTY,
            "frequency_penalty": FREQUENCY_PENALTY,
        }

        outputs = requests.post(API_URL, headers=HEADERS, data=json.dumps(data)).json()
        return outputs

    msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(
        fn=bot,
        inputs=[
            chatbot,
            max_new_tokens,
            temperature,
            top_p,
            top_k,
            system_prompt,
        ],
        outputs=chatbot
    )
    submit_button.click(
        user, [msg, chatbot], [msg, chatbot], queue=False
    ).then(
        fn=bot,
        inputs=[
            chatbot,
            max_new_tokens,
            temperature,
            top_p,
            top_k,
            system_prompt,
        ],
        outputs=chatbot
    )


    def delete_prev_fn(
            history: list[tuple[str, str]]) -> tuple[list[tuple[str, str]], str]:
        try:
            message, _ = history.pop()
        except IndexError:
            message = ''
        return history, message or ''


    def display_input(message: str,
                      history: list[tuple[str, str]]) -> list[tuple[str, str]]:
        history.append((message, ''))
        return history

    retry_button.click(
        fn=delete_prev_fn,
        inputs=chatbot,
        outputs=[chatbot, saved_input],
        api_name=False,
        queue=False,
    ).then(
        fn=display_input,
        inputs=[saved_input, chatbot],
        outputs=chatbot,
        api_name=False,
        queue=False,
    ).then(
        fn=bot,
        inputs=[
            chatbot,
            max_new_tokens,
            temperature,
            top_p,
            top_k,
            system_prompt,
        ],
        outputs=chatbot,
    )

    undo_button.click(
        fn=delete_prev_fn,
        inputs=chatbot,
        outputs=[chatbot, saved_input],
        api_name=False,
        queue=False,
    ).then(
        fn=lambda x: x,
        inputs=[saved_input],
        outputs=msg,
        api_name=False,
        queue=False,
    )

    clear.click(lambda: None, None, chatbot, queue=False)

    gr.Markdown(LICENSE)

# demo.queue(concurrency_count=4, max_size=128)
demo.launch()