[torch] the most concise logging User Guide

Most of the online tutorials are very complex , Actually use logging It's simple , Just three lines of code

I use logging For the convenience of debugging , Because the number of output box caches is limited , If you print the output into a file and observe the use experience, it will be much better , Because the function of these three lines of code is to export the output to a file

The configuration code is as follows ：

import logging

logging.basicConfig(level=logging.INFO, filename='./mylog/test.log', filemode='w')

Here we first quote logging My bag , And then configure it logging Basic configuration

filename Indicates the location of the output file , Note that here he cannot create a new folder , It means... Here 'mylog' You need to create this directory first .

filemode='w' Indicates the mode of modifying the file , This means that every time you run the file, you write the file again, overwriting the content written in the last run , For more file reading and writing modes, please refer to ：  Read write mode

So that's configured , Then basically use it like print equally ：

logging.info('aaaa')

If you want to output variables ： Use ‘f’ Operator ：

aa = 1

logging.info(f'aa:  {aa}')

Just enclose the variables to be output with curly braces , Finally, these outputs are stored in mylog The folder test.log In file . f Please refer to ：https://blog.csdn.net/lwgkzl/article/details/107695293

 

above