【记录】初次本地搭建的模型-MiniCPM 2B

前言

        查阅众多开源大模型后，打算动手尝试搭建端侧模型，看看效果。选中MiniCPM主要是因为参数小，同时中文支持相对较好。

        首先对按照官网提供的demo进行了尝试，然后在colab中完成了一个webui程序并测试，最后通过docker环境在本地搭建并测试成功。

        

Colab Demo测试

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
torch.manual_seed(0)

path = # model path 
tokenizer = AutoTokenizer.from_pretrained(path)
model = AutoModelForCausalLM.from_pretrained(path, torch_dtype=torch.float32, device_map='cuda', trust_remote_code=True)

responds, history = model.chat(tokenizer, "山东省最高的山是哪座山, 它比黄山高还是矮？差距多少？", temperature=0.8, top_p=0.8)
print(responds)

编写程序

from typing import List

import argparse
import gradio as gr
import torch
from threading import Thread
from transformers import (
    AutoModelForCausalLM, 
    AutoTokenizer,
    TextIteratorStreamer
)

import warnings
warnings.filterwarnings('ignore', category=UserWarning, message='TypedStorage is deprecated')

parser = argparse.ArgumentParser()
parser.add_argument("--model_path", type=str, default="")
parser.add_argument("--torch_dtype", type=str, default="bfloat16")
parser.add_argument("--server_name", type=str, default="127.0.0.1")
parser.add_argument("--server_port", type=int, default=7860)

args = parser.parse_args()

# init model torch dtype
torch_dtype = args.torch_dtype
if torch_dtype =="" or torch_dtype == "bfloat16":
    torch_dtype = torch.bfloat16
elif torch_dtype == "float32":
    torch_dtype = torch.float32
else:
    raise ValueError(f"Invalid torch dtype: {torch_dtype}")

# init model and tokenizer
path = args.model_path
tokenizer = AutoTokenizer.from_pretrained(path)
#无显卡auto改为cpu
model = AutoModelForCausalLM.from_pretrained(path, torch_dtype=torch_dtype, device_map="auto", trust_remote_code=True)

# init gradio demo host and port
server_name=args.server_name
server_port=args.server_port

def hf_gen(dialog: List, top_p: float, temperature: float, max_dec_len: int):
    """generate model output with huggingface api

    Args:
        query (str): actual model input.
        top_p (float): only the smallest set of most probable tokens with probabilities that add up to top_p or higher are kept for generation.
        temperature (float): Strictly positive float value used to modulate the logits distribution.
        max_dec_len (int): The maximum numbers of tokens to generate.

    Yields:
        str: real-time generation results of hf model
    """    
    inputs = tokenizer.apply_chat_template(dialog, tokenize=False, add_generation_prompt=False)
    #无显卡去掉.to("cuda")
    enc = tokenizer(inputs, return_tensors="pt").to("cuda")
    streamer = TextIteratorStreamer(tokenizer)
    generation_kwargs = dict(
        enc,
        do_sample=True,
        top_p=top_p,
        temperature=temperature,
        max_new_tokens=max_dec_len,
        pad_token_id=tokenizer.eos_token_id,
        streamer=streamer,
    )
    thread = Thread(target=model.generate, kwargs=generation_kwargs)
    thread.start()
    answer = ""
    for new_text in streamer:
        answer += new_text
        yield answer[4 + len(inputs):]


def generate(chat_history: List, query: str, top_p: float, temperature: float, max_dec_len: int):
    """generate after hitting "submit" button

    Args:
        chat_history (List): [[q_1, a_1], [q_2, a_2], ..., [q_n, a_n]]. list that stores all QA records
        query (str): query of current round
        top_p (float): only the smallest set of most probable tokens with probabilities that add up to top_p or higher are kept for generation.
        temperature (float): strictly positive float value used to modulate the logits distribution.
        max_dec_len (int): The maximum numbers of tokens to generate.

    Yields:
        List: [[q_1, a_1], [q_2, a_2], ..., [q_n, a_n], [q_n+1, a_n+1]]. chat_history + QA of current round.
    """    
    assert query != "", "Input must not be empty!!!"
    # apply chat template
    model_input = []
    for q, a in chat_history:
        model_input.append({"role": "user", "content": q})
        model_input.append({"role": "assistant", "content": a})
    model_input.append({"role": "user", "content": query})
    # yield model generation
    chat_history.append([query, ""])
    for answer in hf_gen(model_input, top_p, temperature, max_dec_len):
        chat_history[-1][1] = answer.strip("</s>")
        yield gr.update(value=""), chat_history


def regenerate(chat_history: List, top_p: float, temperature: float, max_dec_len: int):
    """re-generate the answer of last round's query

    Args:
        chat_history (List): [[q_1, a_1], [q_2, a_2], ..., [q_n, a_n]]. list that stores all QA records
        top_p (float): only the smallest set of most probable tokens with probabilities that add up to top_p or higher are kept for generation.
        temperature (float): strictly positive float value used to modulate the logits distribution.
        max_dec_len (int): The maximum numbers of tokens to generate.

    Yields:
        List: [[q_1, a_1], [q_2, a_2], ..., [q_n, a_n]]. chat_history
    """    
    assert len(chat_history) >= 1, "History is empty. Nothing to regenerate!!"
    # apply chat template
    model_input = []
    for q, a in chat_history[:-1]:
        model_input.append({"role": "user", "content": q})
        model_input.append({"role": "assistant", "content": a})
    model_input.append({"role": "user", "content": chat_history[-1][0]})
    # yield model generation
    for answer in hf_gen(model_input, top_p, temperature, max_dec_len):
        chat_history[-1][1] = answer.strip("</s>")
        yield gr.update(value=""), chat_history


def clear_history():
    """clear all chat history

    Returns:
        List: empty chat history
    """    
    return []


def reverse_last_round(chat_history):
    """reverse last round QA and keep the chat history before

    Args:
        chat_history (List): [[q_1, a_1], [q_2, a_2], ..., [q_n, a_n]]. list that stores all QA records

    Returns:
        List: [[q_1, a_1], [q_2, a_2], ..., [q_n-1, a_n-1]]. chat_history without last round.
    """    
    assert len(chat_history) >= 1, "History is empty. Nothing to reverse!!"
    return chat_history[:-1]


# launch gradio demo
with gr.Blocks(theme="soft") as demo:
    gr.Markdown("""# MiniCPM Gradio Demo""")

    with gr.Row():
        with gr.Column(scale=1):
            top_p = gr.Slider(0, 1, value=0.8, step=0.1, label="top_p")
            temperature = gr.Slider(0.1, 2.0, value=0.8, step=0.1, label="temperature")
            max_dec_len = gr.Slider(1, 1024, value=1024, step=1, label="max_dec_len")
        with gr.Column(scale=5):
            chatbot = gr.Chatbot(bubble_full_width=False, height=400)
            user_input = gr.Textbox(label="User", placeholder="Input your query here!", lines=8)
            with gr.Row():
                submit = gr.Button("Submit")
                clear = gr.Button("Clear")
                regen = gr.Button("Regenerate")
                reverse = gr.Button("Reverse")

    submit.click(generate, inputs=[chatbot, user_input, top_p, temperature, max_dec_len], outputs=[user_input, chatbot])
    regen.click(regenerate, inputs=[chatbot, top_p, temperature, max_dec_len], outputs=[user_input, chatbot])
    clear.click(clear_history, inputs=[], outputs=[chatbot])
    reverse.click(reverse_last_round, inputs=[chatbot], outputs=[chatbot])

demo.queue()
demo.launch(server_name=server_name, server_port=server_port, show_error=True)

Colab程序测试

本地搭建

容器环境

        本地搭建一般个人比较倾向使用 Docker 作为运行环境，在投入很少额外资源的情况下，能够快速获得纯净、可复现的一致性非常棒的环境。

        除此之外，为了高效运行模型，推荐使用 Nvidia 官方的容器镜像（nvcr.io/nvidia/pytorch:24.01-py3[4]）。

        我们可以基于上面的内容，快速搭建一个干净、高效的基础运行环境。

        考虑到我们可能会将模型应用运行在不同的环境，比如云主机和服务器，它们的网络环境可能有所不同。

        当我们本地进行 Docker 镜像构建的时候，配置软件镜像来加速可以大幅改善开发者体验。所以，稍加调整，我们可以得到下面的 Dockerfile 文件：

FROM nvcr.io/nvidia/pytorch:24.01-py3
LABEL maintainer="554686223@qq.com"

# setup Ubuntu and PyPi mirrors, refs: https://github.com/soulteary/docker-stable-diffusion-webui/blob/main/docker/Dockerfile.base
ARG USE_CHINA_MIRROR=true
RUN if [ "$USE_CHINA_MIRROR" = "true" ]; then \
        pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple && \
        sed -i 's/security.ubuntu.com/mirrors.tuna.tsinghua.edu.cn/g' /etc/apt/sources.list && \
        sed -i 's/archive.ubuntu.com/mirrors.tuna.tsinghua.edu.cn/g' /etc/apt/sources.list; \
    fi

# install dependencies、

RUN pip install torch==2.2.1 -i https://pypi.tuna.tsinghua.edu.cn/simple

RUN pip install -i https://pypi.tuna.tsinghua.edu.cn/simple transformers==4.37.2 gradio==4.16.0 accelerate==0.26.1

RUN pip uninstall transformer-engine

将上面的内容保存为 Dockerfile，然后执行下面的命令，可以进行镜像构建：

docker build -t my-gpt -f=dockerfile --no-cache .

下载模型

        根据自身网络情况，选择HuggingFace、ModelScope、WiseModel中最适合你的模型下载或者在线推理平台。

        这里选择了wisemodel，git链接 

         git clone https://www.wisemodel.cn/OpenBMB/miniCPM-dpo-fp32.git

简单测试

启动容器

docker run --rm -it -p 7860:7860 --gpus all --ipc=host --ulimit memlock=-1 -v D:/weiyisoftware/gpttest/gptdocker/models:/app/models -v D:/weiyisoftware/gpttest/gptdocker/workspace:/workspace my-gpt python app.py --model_path=/app/models/OpenBMB/miniCPM-dpo-fp32/ --server_name=0.0.0.0 --torch_dtype=float32

成功后：http://localhost:7860。

ps：因为电脑配置问题，回复很慢。