[1] Visual interface

   First on the renderings ：

[2] Use the tutorial

[2.1] install pytorch and tensorboard

   Enter the command ：

pip install torch
pip install tensorboard 

[2.2] Used in code write

   Code demo：

import numpy as np
from torch.utils.tensorboard import SummaryWriter  #  You can also use  tensorboardX
# from tensorboardX import SummaryWriter #  You can also use  pytorch  Integrated  tensorboard

writer = SummaryWriter('log') #  Configure the address where the generated data is saved 
for epoch in range(100):
    writer.add_scalar('test/squared', np.square(epoch), epoch)

writer.close()

   After executing the above code, a logs file , And contains an event file .

[2.3] open tensorboard panel

   stay pycharm in terminal The input terminal ：

tensorboard --logdir=logs

   Be sure to pay attention to the initial configuration of the generated file storage address , you are here terminal The address of the command in the terminal should be accessible ！！！

   After typing the command , Will generate an address , Just visit .

   Pay attention to using Google browser to access . I use QQ browser （ Google kernel ） Some data can't be printed .