置顶，懒得看具体细节的同学可以直接用 我的GitHub安装脚本 进行环境配置

1 环境配置

1.1 工作环境

自带 python3.6 + python3.6-dev + TensorFlow2.0 + TF2.0相关包

1.2 其他美化

zsh + oh_my_zsh 美化终端 ； SpaceVim 加强vim

2 常见问题

2.1 APT 包找不到

建议使用 中科大 源， 而非 清华 源，因为清华源对 Ubuntu 和 Ros 更新速度较慢。
（如本次我就没有在清华源找到 python3.6-dev ）

2.2 TF 编译附带包失败

Nvidia 官方 TF for JetPack 讲的问题比较清楚，但缺少了安装 python3.6-dev 。
缺少dev会导致cython在编译包时失败并提示 unable to find pgen, not compiling formal grammar. 这是一个非常有误导性的错误，我尝试安装了对应版本的pgen，但仍然报错；直到在StackOverFlow 上找到一个相似的问题，安装dev后解决

2.3 TF包下载慢

我的CSDN下载链接 ，不需要积分

3 代码

#！/bin/bash
echo -e "\033[1;31m Install basic env\033[0m"
echo -e "\033[1;32m Changing APT.Sources...\033[0m"
sleep 2

echo "# Update APT-Source for aarch64
# 默认注释了源码仓库，如有需要可自行取消注释
deb https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic main restricted universe multiverse
# deb-src https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic main main restricted universe multiverse
deb https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic-updates main restricted universe multiverse
# deb-src https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic-updates main restricted universe multiverse
deb https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic-backports main restricted universe multiverse
# deb-src https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic-backports main restricted universe multiverse
deb https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic-security main restricted universe multiverse
# deb-src https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic-security main restricted universe multiverse
# 预发布软件源，不建议启用
# deb https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic-proposed main restricted universe multiverse
# deb-src https://mirrors.ustc.edu.cn/ubuntu-ports/ bionic-proposed main restricted universe multiverse
" > sources.list

echo -e "\033[1;32m\n Updating System \n...\033[0m"
sleep 2

sudo mv /etc/apt/sources.list /etc/apt/sources.list.bak
sudo mv sources.list /etc/apt/sources.list

sleep 2

sudo apt update
sudo apt update
sudo apt -y uprade
sudo apt install -y python3-pip python3.6-dev libcurl3-gnutls=7.47.0-1ubuntu2.14 zsh curl tmux screen tree
sudo apt -y autoremove

echo -e "\033[1;32m\n Export CUDA PATH \n...\033[0m"
echo -e "
export CUDA_HOME=/usr/local/cuda-10.0
export LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:$LD_LIBRARY_PATH
export PATH=/usr/local/cuda-10.0/bin:$PATH
" >> ~/.bashrc

echo -e "\033[1;32m\n SpaceVIM \n...\033[0m"
curl -sLf https://spacevim.org/cn/install.sh | bash


echo -e "\033[1;32m\n oh-my-zsh \n...\033[0m"
curl -fsSL https://raw.githubusercontent.com/ohmyzsh/ohmyzsh/master/tools/install.sh | bash
chsh -s /usr/bin/zsh

echo -e "\033[1;32m\n install TF-2.0 \n...\033[0m"
wget https://developer.download.nvidia.com/compute/redist/jp/v42/tensorflow-gpu/tensorflow_gpu-2.0.0+nv19.11-cp36-cp36m-linux_aarch64.whl

sudo apt-get install libhdf5-serial-dev hdf5-tools libhdf5-dev zlib1g-dev zip libjpeg8-dev
sudo pip3 install -U pip testresources setuptools

sudo pip3 install -U tensorflow_gpu-2.0.0+nv19.11-cp36-cp36m-linux_aarch64.whl Processing ./tensorflow_gpu-2.0.0+nv19.11-cp36-cp36m-linux_aarch64.whl


# If you don't want to use TF 2.0
# You can try other
#  sudo pip3 install --pre --extra-index-url https://developer.download.nvidia.com/compute/redist/jp/v43 tensorflow-gpu