山东大学项目实训纪实 2024.5.11

使用vicuna7b-v0跑出来的模型，根据这个跑出来的模型，写一个小的demo展示一下，后面再完善ui和管理端

写出来的小demo大致长这个样子（简单一写，等回头写前后端的同学写好了ui再合起来）

目前可以做到给一个smiles，进行提问，回答该smiles的一些特点之类的。（但是vicuna对中文来说不太友好，建议提问的时候用英文提问会更精确一些，因为训练编的QA也是英文的）

目前有个缺点，就是7b的模型，精度不高，回答的问题有些地方有错误，还需要再调一下

import argparse
import os
import random

import numpy as np
import torch
import torch.backends.cudnn as cudnn
import gradio as gr

from pipeline.common.config import Config
from pipeline.common.dist_utils import get_rank
from pipeline.common.registry import registry
from pipeline.conversation.conversation import Chat, CONV_VISION

# imports modules for registration
from pipeline.datasets.builders import *
from pipeline.models import *
from pipeline.processors import *
from pipeline.runners import *
from pipeline.tasks import *


#定义命令参数解析函数
def parse_args():
    parser = argparse.ArgumentParser(description="Demo")
    parser.add_argument("--cfg-path", required=True, help="path to configuration file.")
    parser.add_argument("--gpu-id", type=int, default=0, help="specify the gpu to load the model.")
    parser.add_argument(
        "--options",
        nargs="+",
        help="override some settings in the used config, the key-value pair "
        "in xxx=yyy format will be merged into config file (deprecate), "
        "change to --cfg-options instead.",
    )
    args = parser.parse_args()
    return args

#随机种子
def setup_seeds(config):
    seed = config.run_cfg.seed + get_rank()

    random.seed(seed)
    np.random.seed(seed)
    torch.manual_seed(seed)

    cudnn.benchmark = False
    cudnn.deterministic = True


# ========================================
#             Model Initialization
# ========================================

#初始化chat
print('Initializing Chat')
args = parse_args()
cfg = Config(args)
use_amp = cfg.run_cfg.get("amp", False)

model_config = cfg.model_cfg
model_config.device_8bit = args.gpu_id
model_cls = registry.get_model_class(model_config.arch)
print(model_config)
model = model_cls.from_config(model_config)

model = model.to('cuda:{}'.format(args.gpu_id))

vis_processor_cfg = cfg.datasets_cfg.cc_sbu_align.vis_processor.train
vis_processor = registry.get_processor_class(vis_processor_cfg.name).from_config(vis_processor_cfg)
chat = Chat(model, vis_processor, device='cuda:{}'.format(args.gpu_id))
print('Initialization Finished')

# ========================================
#             Gradio Setting
# ========================================

def gradio_reset(chat_state, img_list):
    if chat_state is not None:
        chat_state.messages = []
    if img_list is not None:
        img_list = []
    return None, gr.update(value=None, interactive=True), gr.update(placeholder='Please input your SMILES string first', interactive=False),gr.update(value="Input SMILES & Start Chat", interactive=True), chat_state, img_list

@torch.no_grad()
def upload_img(gr_img, text_input, chat_state):
    if gr_img is None:
        return None, None, gr.update(interactive=True), chat_state, None
    chat_state = CONV_VISION.copy()
    img_list = []
    print(gr_img)
    print(chat_state)
    with torch.cuda.amp.autocast(use_amp):
        llm_message = chat.upload_img(gr_img, chat_state, img_list)
    return gr.update(interactive=False), gr.update(interactive=True, placeholder='Type and press Enter'), gr.update(value="Start Chatting", interactive=False), chat_state, img_list

def gradio_ask(user_message, chatbot, chat_state):
    if len(user_message) == 0:
        return gr.update(interactive=True, placeholder='Input should not be empty!'), chatbot, chat_state
    chat.ask(user_message, chat_state)
    chatbot = chatbot + [[user_message, None]]
    return '', chatbot, chat_state


@torch.no_grad()
def gradio_answer(chatbot, chat_state, img_list, num_beams, temperature):
    with torch.cuda.amp.autocast(use_amp):
        llm_message = chat.answer(conv=chat_state,
                                img_list=img_list,
                                num_beams=num_beams,
                                temperature=temperature,
                                max_new_tokens=300,
                                max_length=2000)[0]
    chatbot[-1][1] = llm_message
    return chatbot, chat_state, img_list

title = """<h1 align="center">Demo of DrugTeller</h1>"""
description = """<h3>This is the demo of DrugTeller. Input your SMILES string and start chatting!</h3>"""


#TODO show examples below

with gr.Blocks() as demo:
    gr.Markdown(title)
    gr.Markdown(description)

    with gr.Row():
        with gr.Column(scale=0.5):
            smiles = gr.Textbox(label="SMILES")
            upload_button = gr.Button(value="Input SMILES & Start Chat", interactive=True, variant="primary")
            clear = gr.Button("Restart")
            
            num_beams = gr.Slider(
                minimum=1,
                maximum=10,
                value=1,
                step=1,
                interactive=True,
                label="Beam search numbers",
            )
            
            temperature = gr.Slider(
                minimum=0.1,
                maximum=2.0,
                value=1.0,
                step=0.1,
                interactive=True,
                label="Temperature",
            )

        with gr.Column():
            chat_state = gr.State()
            img_list = gr.State()
            chatbot = gr.Chatbot(label='DrugTeller')
            text_input = gr.Textbox(label='User', placeholder='Please input your SMILES string first', interactive=False)
    
    upload_button.click(upload_img, [smiles, text_input, chat_state], [smiles, text_input, upload_button, chat_state, img_list])
    
    text_input.submit(gradio_ask, [text_input, chatbot, chat_state], [text_input, chatbot, chat_state]).then(
        gradio_answer, [chatbot, chat_state, img_list, num_beams, temperature], [chatbot, chat_state, img_list]
    )
    clear.click(gradio_reset, [chat_state, img_list], [chatbot, smiles, text_input, upload_button, chat_state, img_list], queue=False)

demo.launch(share=True, enable_queue=True)