一、基础环境搭建
1、python3 
准备相关的jar包  Index of /ftp/python/3.7.9/

scp Python-3.7.9.tgz root@192.168.1.245:/opt/dockerinstall/python3/

yum -y install gcc

yum -y install zlib-devel bzip2-devel openssl-devel ncurses-devel sqlite-devel readline-devel tk-devel gdbm-devel db4-devel libpcap-devel xz-devel libffi-devel

tar -zxvf Python-3.7.9.tgz

mkdir /usr/local/python3

cd Python-3.7.9/

./configure --prefix=/usr/local/python3

make && make install

sudo ln -sf /usr/local/python3/bin/python3.7 /usr/bin/python3

sudo ln -sf /usr/local/python3/bin/pip3.7 /usr/bin/pip3

配置python3的路径

cat ~/.bashrc

vim ~/.bashrc 

添加

export PATH="/usr/local/python3/bin:$PATH"

source ~/.bashrc 

2、安装 paddlenlp 2.5.2

升级pip3 

/usr/local/python3/bin/python3.7 -m pip install --upgrade pip

参考 飞桨PaddlePaddle-源于产业实践的开源深度学习平台

pip3 install paddlenlp==2.5.2 -i https://pypi.tuna.tsinghua.edu.cn/simple

代码安装
参考：PaddleNLP/docs/get_started/installation.rst at develop · PaddlePaddle/PaddleNLP · GitHub

git clone https://github.com/PaddlePaddle/PaddleNLP.git  下载不了，用之前已下载的工程

多试几次 下载成功

最新的develop分支 没有applications目录，还是需要用之前已下载的工程

scp root@192.168.1.243:/opt/dockerinstall/PaddleNLP.tar.gz .

cd /opt/dockerinstall/PaddleNLP/applications/text_classification/multi_class/deploy/simple_serving

pip3 install paddlepaddle -i https://pypi.tuna.tsinghua.edu.cn/simple

启动命令

nohup python3 -m uvicorn main_new:app --host 192.168.1.245 --port 6001 --workers 4 > paddlenlpweb_new.log 2>&1 &