app.py

model_name = "berkeley-nest/Starling-LM-7B-alpha"

title = """# 👋🏻Welcome to Tonic's 💫🌠Starling 7B"""
description = """You can use [💫🌠Starling 7B](https://huggingface.co/berkeley-nest/Starling-LM-7B-alpha) or duplicate it for local use or on Hugging Face! [Join me on Discord to build together](https://discord.gg/VqTxc76K3u)."""

import transformers
from transformers import AutoConfig, AutoTokenizer, AutoModel, AutoModelForCausalLM
import torch
import gradio as gr
import json
import os
import shutil
import requests
import accelerate
import bitsandbytes
import gc

device = "cuda" if torch.cuda.is_available() else "cpu"
bos_token_id = 1,
eos_token_id = 32000
pad_token_id = 32001
temperature=0.4
max_new_tokens=240
top_p=0.92
repetition_penalty=1.7

tokenizer = transformers.AutoTokenizer.from_pretrained(model_name)
model =  AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
model.eval()
os.environ['PYTORCH_CUDA_ALLOC_CONF'] = 'max_split_size_mb:50'

class StarlingBot:
    def __init__(self, assistant_message="I am Starling-7B by Tonic-AI, I am ready to do anything to help my user."):
        self.assistant_message = assistant_message

    def predict(self, user_message, assistant_message, mode, do_sample, temperature=0.4, max_new_tokens=700, top_p=0.99, repetition_penalty=1.9):
        try:
            if mode == "Assistant":
                conversation = f"GPT4 Correct Assistant: {assistant_message if assistant_message else ''}  GPT4 Correct User: {user_message}  GPT4 Correct Assistant:"
            else:  # mode == "Coder"
                conversation = f"Code Assistant: {assistant_message if assistant_message else ''}  Code User:: {user_message}  Code Assistant:"
            input_ids = tokenizer.encode(conversation, return_tensors="pt", add_special_tokens=True)
            input_ids = input_ids.to(device)
            response = model.generate(
                input_ids=input_ids, 
                use_cache=True, 
                early_stopping=False, 
                bos_token_id=bos_token_id, 
                eos_token_id=eos_token_id, 
                pad_token_id=pad_token_id, 
                temperature=temperature, 
                do_sample=True, 
                max_new_tokens=max_new_tokens, 
                top_p=top_p, 
                repetition_penalty=repetition_penalty
            )
            response_text = tokenizer.decode(response[0], skip_special_tokens=True)
#           response_text = response.split("<|assistant|>\n")[-1]
            return response_text       
        finally:
            del input_ids
            gc.collect()
            torch.cuda.empty_cache()

examples = [
    [
        "The following dialogue is a conversation between Emmanuel Macron and Elon Musk:",  # user_message
        "[Emmanuel Macron]: Hello Mr. Musk. Thank you for receiving me today.",  # assistant_message
        0.9,  # temperature
        450,  # max_new_tokens
        0.90,  # top_p
        1.9,  # repetition_penalty
    ]
]

starling_bot = StarlingBot()

def gradio_starling(user_message, assistant_message, mode, do_sample, temperature, max_new_tokens, top_p, repetition_penalty):
    response = starling_bot.predict(user_message, assistant_message, mode, do_sample, temperature, max_new_tokens, top_p, repetition_penalty)
    return response

with gr.Blocks(theme="ParityError/Anime") as demo:
    gr.Markdown(title)  
    gr.Markdown(description)  
    with gr.Row():
        assistant_message = gr.Textbox(label="Optional💫🌠Starling Assistant Message", lines=2)
        user_message = gr.Textbox(label="Your Message", lines=3)
    with gr.Row():
        mode = gr.Radio(choices=["Assistant", "Coder"], value="Assistant", label="Mode")
        do_sample = gr.Checkbox(label="Advanced", value=True)    
    with gr.Accordion("Advanced Settings", open=lambda do_sample: do_sample):
        with gr.Row():
            temperature = gr.Slider(label="Temperature", value=0.4, minimum=0.05, maximum=1.0, step=0.05)
            max_new_tokens = gr.Slider(label="Max new tokens", value=100, minimum=25, maximum=800, step=1)
            top_p = gr.Slider(label="Top-p (nucleus sampling)", value=3.6, minimum=1.0, maximum=4.0, step=0.1)
            repetition_penalty = gr.Slider(label="Repetition penalty", value=1.9, minimum=1.0, maximum=2.0, step=0.05)

    submit_button = gr.Button("Submit")
    output_text = gr.Textbox(label="💫🌠Starling Response")

    submit_button.click(
        gradio_starling,
        inputs=[user_message, assistant_message, mode, do_sample, temperature, max_new_tokens, top_p, repetition_penalty],
        outputs=output_text
    )

demo.launch()