深度学习环境搭建笔记Ubuntu 16.04 RTX2080Ti Cuda 10.0 Cudnn 7.4 Python 3.6(anaconda) Tensorflow 1.13.1 (附有用链接）
最近折腾着配置了几次深度学习的环境，把个人觉得有用的链接附上，大家可以做个参考，欢迎大佬们批评指正。因为大家估计也免不了反复折腾，所以卸载和安装的教程链接一同奉上了。
1. ubuntu 16.04的安装
- ubuntu 16.04的启动U盘制作
https://blog.csdn.net/fu6543210/article/details/79722615
- ubuntu 16.04的安装
https://blog.csdn.net/ramse/article/details/52716937?utm_medium=distribute.pc_relevant_t0.none-task-blog-BlogCommendFromMachineLearnPai2-1.nonecase&depth_1-utm_source=distribute.pc_relevant_t0.none-task-blog-BlogCommendFromMachineLearnPai2-1.nonecase
- 磁盘分区（个人认为最难的部分）
https://askubuntu.com/questions/343268/how-to-use-manual-partitioning-during-installation
http://www.imooc.com/read/39/article/458
https://zhuanlan.zhihu.com/p/35294113
https://askubuntu.com/questions/343268/how-to-use-manual-partitioning-during-installation
https://blog.csdn.net/ctx_ubuntu/article/details/90693617?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522159271683919725250105324%2522%252C%2522scm%2522%253A%252220140713.130102334.pc%255Fall.%2522%257D&request_id=159271683919725250105324&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2allfirst_rank_v2~rank_v25-3-90693617.first_rank_v2_rank_v25&utm_term=ubuntu%E5%8D%95%E7%B3%BB%E7%BB%9F%E5%AE%89%E8%A3%85%E7%A1%AC%E7%9B%98%E5%88%86%E5%8C%BA
**
2.Nvidia驱动的安装与卸载:
**
- 卸载
https://blog.csdn.net/qq_28612967/article/details/102787717?utm_medium=distribute.pc_relevant.none-task-blog-BlogCommendFromMachineLearnPai2-1.nonecase&depth_1-utm_source=distribute.pc_relevant.none-task-blog-BlogCommendFromMachineLearnPai2-1.nonecase
如果驱动很新的话，可能会无法用上面文章中的指令卸载，这个时候可以用下面文章的办法。
https://blog.csdn.net/ezhchai/article/details/80536949
- 安装
推荐用下面这种方法，虽然麻烦一些，但是整个过程更可控。我用那种简单的方法安装时，出现了循环登录。
https://blog.csdn.net/sinat_42239797/article/details/101618334
注：如果是单显卡的情况，运行下面的指令，会导致opengl files不安装。
sudo ./NVIDIA-Linux-x86_64-430.50 .run -no-x-check -no-nouveau-check -no-opengl-files
这个可能会导致图形界面卡顿，所以推荐单显卡的话去掉最后那个命令，运行下面的指令
sudo ./NVIDIA-Linux-x86_64-430.50 .run -no-x-check -no-nouveau-check
容易的方法
https://blog.csdn.net/qq_35839335/article/details/102471087?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522159318846419724811834303%2522%252C%2522scm%2522%253A%252220140713.130102334.pc%255Fall.%2522%257D&request_id=159318846419724811834303&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2allfirst_rank_ecpm_v1~rank_ctr_v4-5-102471087.ecpm_v1_rank_ctr_v4&utm_term=RTX2080Ti%E6%8E%A8%E8%8D%90%E7%9A%84%E7%8E%AF%E5%A2%83%E9%85%8D%E7%BD%AE
3.cuda and cudnn的安装与卸载
- 安装
因为我用的是RTX2080Ti, 所以推荐安装Cuda10.0, Cudnn 7.4， tensorflow 1.13.1。最开始装cuda8.0， cudnn6.0，tensorflow 1.4.0, 运行深度学习模型会有无法启用gpu的情况。
https://blog.csdn.net/cskywit/article/details/97239371
https://blog.csdn.net/IT_xiao_bai/article/details/84782311
- 卸载
https://blog.csdn.net/hjxu2016/article/details/97023248
https://blog.csdn.net/wanzhen4330/article/details/81704474?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522159322097019724846407750%2522%252C%2522scm%2522%253A%252220140713.130102334…%2522%257D&request_id=159322097019724846407750&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2allfirst_rank_ecpm_v1~rank_ctr_v4-4-81704474.ecpm_v1_rank_ctr_v4&utm_term=%E5%8D%B8%E8%BD%BDcudnn
4. gcc的安装与卸载
因为要与所安装的cuda版本匹配，所以有时需要对gcc进行升级或者降级。
这种方法十分简单，ubuntu16.04亲测有效。
https://blog.csdn.n
et/liuweizj12/article/details/90548843?utm_medium=distribute.pc_relevant.none-task-blog-baidujs-2
下面的方法可以把gcc升级或降级到任意版本
https://blog.csdn.net/flowerspring/article/details/88885050
4. Anaconda的安装与卸载
Anaconda, tensorflow和pytorch的安装与卸载相对比较简单
https://blog.csdn.net/qq_41271957/article/details/90111862?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522159299675619724835814519%2522%252C%2522scm%2522%253A%252220140713.130102334.pc%255Fall.%2522%257D&request_id=159299675619724835814519&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2allfirst_rank_ecpm_v1~rank_business_v1-4-90111862.ecpm_v1_rank_business_v1&utm_term=ubuntu+16.04%E9%85%8D%E7%BD%AE%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%9A%84%E7%8E%AF%E5%A2%83
https://blog.csdn.net/channing_q/article/details/86564750?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522159318846419724811834303%2522%252C%2522scm%2522%253A%252220140713.130102334.pc%255Fall.%2522%257D&request_id=159318846419724811834303&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2allfirst_rank_ecpm_v1~rank_ctr_v4-2-86564750.ecpm_v1_rank_ctr_v4&utm_term=RTX2080Ti%E6%8E%A8%E8%8D%90%E7%9A%84%E7%8E%AF%E5%A2%83%E9%85%8D%E7%BD%AE
记得给anaconda导入path
5. Tensorflow的安装和卸载
官方推荐的环境配置列表
https://tensorflow.google.cn/install/source
- 安装：
pip install tensorflow-gpu==version
可以直接用pip，会自动安装到anaconda的路径里
- 卸载
https://blog.csdn.net/greenlight_74110/article/details/78408554
https://blog.csdn.net/FANGUOHAO/article/details/88411352
6.pytorch的安装与卸载
https://blog.csdn.net/miao0967020148/article/details/80400357
https://pytorch.org/get-started/previous-versions/
总结
这些链接是我最近这几天反复安装卸载几次积累下来的，感觉坑比较少的一些操作步骤。大家如果遇到什么坑，欢迎多多交流！