书生大模型-基础岛2-8G 显存玩转书生大模型 Demo

任务详情: https://github.com/InternLM/Tutorial/blob/camp3/docs/L1/Demo/task.md

基础任务（完成此任务即完成闯关）

使用 Cli Demo 完成 InternLM2-Chat-1.8B 模型的部署，并生成 300 字小故事，记录复现过程并截图。

step1：配置环境

# 创建环境
conda create -n demo python=3.10 -y
# 激活环境
conda activate demo
# 安装 torch
conda install pytorch==2.1.2 torchvision==0.16.2 torchaudio==2.1.2 pytorch-cuda=12.1 -c pytorch -c nvidia -y
# 安装其他依赖
pip install transformers==4.38
pip install sentencepiece==0.1.99
pip install einops==0.8.0
pip install protobuf==5.27.2
pip install accelerate==0.33.0
pip install streamlit==1.37.0

这里使用的是上海AI Lab，提供的服务器，可以直接clone开盒即用的环境：

 (base) root@intern-studio-40077097:~# studio-conda -h
  说明: 用于快速clone预设的conda环境
  使用: 
    1. studio-conda env -l/list 打印预设的conda环境列表
    2. studio-conda <target-conda-name> 快速clone: 默认拷贝internlm-base conda环境
    3. studio-conda -t <target-conda-name> -o <origin-conda-name> 将预设的conda环境拷贝到指定的conda环境

step2：运行脚本

源码来自官方教程文档

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM


model_name_or_path = "/root/share/new_models/Shanghai_AI_Laboratory/internlm2-chat-1_8b"

tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, trust_remote_code=True, device_map='cuda:0')
model = AutoModelForCausalLM.from_pretrained(model_name_or_path, trust_remote_code=True, torch_dtype=torch.bfloat16, device_map='cuda:0')
model = model.eval()

system_prompt = """You are an AI assistant whose name is InternLM (书生·浦语).
- InternLM (书生·浦语) is a conversational language model that is developed by Shanghai AI Laboratory (上海人工智能实验室). It is designed to be helpful, honest, and harmless.
- InternLM (书生·浦语) can understand and communicate fluently in the language chosen by the user such as English and 中文.
"""

messages = [(system_prompt, '')]

print("=============Welcome to InternLM chatbot, type 'exit' to exit.=============")

while True:
    input_text = input("\nUser  >>> ")
    input_text = input_text.replace(' ', '')
    if input_text == "exit":
        break

    length = 0
    for response, _ in model.stream_chat(tokenizer, input_text, messages):
        if response is not None:
            print(response[length:], flush=True, end="")
            length = len(response)

进阶任务（闯关不要求完成此任务）

P1. 使用 LMDeploy 完成 InternLM-XComposer2-VL-1.8B 的部署，并完成一次图文理解对话，记录复现过程并截图。

复现步骤：

环境：cuda12+10% A100

step1 激活环境并安装 LMDeploy 以及其他依赖。

conda activate demo
pip install lmdeploy[all]==0.5.1
pip install timm==1.0.7

step2 使用 LMDeploy 启动一个与 InternLM-XComposer2-VL-1.8B 模型交互的 Gradio 服务。

lmdeploy serve gradio /share/new_models/Shanghai_AI_Laboratory/internlm-xcomposer2-vl-1_8b --cache-max-entry-count 0.1

step3 在使用 Upload Image 上传图片后，输入 Instruction 后按下回车，便可以看到模型的输出。

可能会遇到的问题：

(1) 找不到 frpc_linux_amd64_v0.2

Please check your internet connection. This can happen if your antivirus software blocks the download of this file. You can install manually by following these steps: 

1. Download this file: https://cdn-media.huggingface.co/frpc-gradio-0.2/frpc_linux_amd64
2. Rename the downloaded file to: frpc_linux_amd64_v0.2
3. Move the file to this location:llama3/lib/python3.9/site-packages/gradio

解决：下载frpc_linux_amd64，更名为frpc_linux_amd64_v0.2，移动到本虚拟环境的指定目录下mv frpc_linux_amd64_v0.2 .....python3.9/site-packages/gradio

（2）gradio报错：

Could not create share link. Please check your internet connection or our status page: https://status.gradio.app

解决：cd进入frpc_linux_amd64文件的位置（应该在.conda/envs/ChatGLM2/lib/python3.11/site-packages/gradio），输入以下命令给予权限：chmod +x frpc_linux_amd64_v0.2

参考：
[1] https://blog.csdn.net/MMMMMMNONG/article/details/140128052
[2] https://blog.csdn.net/powerbond/article/details/132911392

P2. 使用 LMDeploy 完成 InternVL2-2B 的部署，并完成一次图文理解对话，记录复现过程并截图。

复现步骤

step1 激活环境并启动gradio

conda activate demo
lmdeploy serve gradio /share/new_models/OpenGVLab/InternVL2-2B --cache-max-entry-count 0.1

step2 上传图片，输入指令