Table of Contents
6、配置ssh的协议 + pycharm + xshell 连接服务器
引言
该教程的准备前,需要在服务器上安装docker
- 其他的一些参考资料
pycharm 远程连接docker容器调试程序
https://blog.csdn.net/hanchaobiao/article/details/84069299
PyCharm+Docker:打造最舒适的深度学习炼丹炉
https://zhuanlan.zhihu.com/p/52827335
docker的创建:
1、服务器上安装好docker
2、查找安装对应的镜像
在下面的docker hub 中找到自己合适的镜像
Docker Hub
如,下载一个pytorch相关的镜像
Explore - Docker Hub
https://hub.docker.com/search?q=pytorch&type=image
3、在终端中下载相关的镜像
docker pull dsksd/pytorch:0.4.1
docker pull 下载镜像的命令
dsksd/pytorch 为镜像的名称
0.4.1 为对应的标签的版本
4、docker images 查看相关的镜像的 id
xyj@xyj-W580-G20:~$ docker images
WARNING: Error loading config file: /home/xyj/.docker/config.json: stat /home/xyj/.docker/config.json: permission denied
REPOSITORY TAG IMAGE ID CREATED SIZE
nvidia/cuda 9.0-base 1476e7f683da 2 weeks ago 175MB
nvidia/cuda 9.0-base-ubuntu16.04 cfab853500aa 2 weeks ago 174MB
nvidia/cuda 10.0-base 1ab2d3e6e58d 2 weeks ago 112MB
mycaffe/gpu 1.0 6570aaff012f 2 weeks ago 5.17GB
nvidia/cuda <none> 721bf18dc98f 5 weeks ago 137MB
nvidia/cuda <none> e5eb52810532 5 weeks ago 110MB
caffe/gpu 0.1 8deef48b1da2 5 weeks ago 15.8GB
caffe 0.1 697b3f6e7043 6 weeks ago 14.4GB
pytorch/pytorch nightly-devel-cuda10.0-cudnn7 a5b30513fbc1 2 months ago 6.4GB
nvidia/cuda <none> cec06c83b6d6 3 months ago 137MB
nvidia/cuda <none> e0719fc05fac 3 months ago 165MB
nvidia/cuda <none> 53418d80213d 3 months ago 115MB
nvidia/cuda 9.0-cudnn7-runtime-ubuntu16.04 3b75042d0c57 3 months ago 1.22GB
nvidia/cuda 10.0-cudnn7-runtime-ubuntu16.04 37aa75fe187f 3 months ago 1.33GB
daocloud.io/daocloud/tensorflow latest c9a0882cbdbc 4 months ago 1.05GB
tensorflow/tensorflow latest c9a0882cbdbc 4 months ago 1.05GB
ufoym/deepo latest 3256880e4c5b 4 months ago 10.6GB
kbobrowski/tensorflow-gpu-opencv latest 38069988e9ae 5 months ago 3.91GB
container-314 latest bfc13bc0d9c9 5 months ago 19.2GB
c-new latest 905996b83494 5 months ago 14.8GB
nvidia/cuda 9.0-cudnn7-devel-ubuntu16.04 f780be4907ca 6 months ago 2.71GB
nvidia/cuda 10.0-cudnn7-devel-ubuntu16.04 fdbd6d128838 6 months ago 3.09GB
nvidia/cuda 8.0-devel a8f4ac5ee686 6 months ago 1.7GB
tensorflow/tensorflow <none> 72bdfab04989 6 months ago 1.1GB
nvidia/cuda latest d9a8427c8dd9 7 months ago 2.35GB
dbctraining/pytorch0.4.1-gpu-cuda9-cudnn7-py3 v1.0.1 f57cdd7c8b7f 7 months ago 3.04GB
anibali/pytorch cuda-9.2 86b1edf0bd95 8 months ago 3.51GB
daocloud.io/daocloud/tensorflow nightly-devel-gpu-py3 beb1f2c06280 9 months ago 3.88GB
blinkeye/deepo-tf-caffe-opencv-dlib-py3 latest 294b5124decf 10 months ago 4.74GB
dsksd/pytorch 0.4.1 a8bdabe9931a 11 months ago 11.7GB
tensorflow/tensorflow 1.8.0-devel-gpu-py3 fff42c2fd81c 16 months ago 3.13GB
bvlc/caffe gpu ba28bcb1294c 16 months ago 3.38GB
rynge/osg-tensorflow-gpu latest e5100c70954f 22 months ago 7.14GB
drunkar/anaconda-tensorflow-gpu latest a120be75d222 3 years ago 3.61GB
xyj@xyj-W580-G20:~$ docker images | grep dsksd
WARNING: Error loading config file: /home/xyj/.docker/config.json: stat /home/xyj/.docker/config.json: permission denied
dsksd/pytorch 0.4.1 a8bdabe9931a 11 months ago 11.7GB
docker images | grep dsksd
| 管道符,同时运行两个命令
grep 过滤
dsksd 过滤要找的关键词
5、创建容器
xyj@xyj-W580-G20:~$ sudo docker run --name xjr_pytocrh -it -v /home/xyj/xjr_directory/:/home/xjr_directory -p 6898:22 -p 6806:6066 -p 6088:8888 a8bda /bin/bash
说明:
-v : 目录映射
/home/xyj/xjr_directory/ : 本机目录
/home/xjr_directory :docker 容器里面的目录
-p :端口映射
4、5 相关的截图
6、配置ssh的协议 + pycharm + xshell 连接服务器
pycharm
映射路径说明:
Root path:pycharm 直接连接的服务器的目录
Deployment path on server : 相对于 Root path 的服务器目录
下图中由于 Root path 是 / ,因此 Deployment path on server 就是服务器的绝对目录
等价于:
Root path:
/home/docker/code
Deployment path on server:
./app
pycharm 远程连接docker容器调试程序
https://blog.csdn.net/hanchaobiao/article/details/84069299
PyCharm+Docker:打造最舒适的深度学习炼丹炉
https://zhuanlan.zhihu.com/p/52827335
Xshell
(1条消息)Xshell如何连接Docker容器中的Linux - u010046887的专栏 - CSDN博客
https://blog.csdn.net/u010046887/article/details/90406725
7、退出容器后,重新启动该容器并且在后台运行
- 退出容器
exit
- 找到该容器
xyj@xyj-W580-G20:~$ docker ps
- 启动容器
9ce 容器id号码
xyj@xyj-W580-G20:~$ docker start 9ce
- 重新启动该容器,并在后台运行
xyj@xyj-W580-G20:~$ docker exec -it 9ce bash