Article Source| Hengyuan Cloud Community
Original address| Automated background training
Author | Littleor
Hello, my editor is still working hard today, not crazy. Although I was told that the area I am in is a closed area, I still need to complete the work. I have to work hard! ⛽️
Today, I will not carry the thesis articles of my friends, but I will share with you a skill first!
The content is not long, and I hope it can give you a little help for alchemy friends.
Here is a simple and easy-to-use script for fully automated training log recording in the background, automatic saving of result data and automatic shutdown.
Warehouse address: https://github.com/Littleor/HY-Auto-Trainer
characteristic
- Use it to train
nohup
in the background, fullyssh
disconnect and keep your hands free. OSS
After the training is completed, the running results are automatically packaged to avoid loss of results.- Automatic shutdown after uploading, saving money and trouble~
usage
configuration
- Copy all *.sh files in the directory to the project directory (such as /hy-tmp)
- Add permissions to *.sh files:
chmod a+x *.sh
- Modify the logPath in run.sh to modify the log save path
- Modify the resultPath in end.sh to modify the path of the running results that need to be saved
- If it is not a Python project, you need to modify the python of end.sh to the corresponding execution command
use
Taking this project as an example, there is a Train/main.py file in the project as the simulation training file, and the project file is placed in the /hy-tmp directory.
How to run:
# ./run.sh [可执行文件目录] [Tag 用于标注结果]
./run.sh /hy-tmp/Train/main.py Test-Trainer