For detailed steps, please see: https://blog.csdn.net/index20001/article/details/82871634
Since my machine can connect to the Linux server normally through SSH, and Tensorboard can be used normally on the server, so I cut directly to the topic and forwarded the server port to this machine
1. Open the Xshell client, open the target session window, click "Connection" → "SSH" → "Tunnel", and add "Transfer Rules", as follows:
2. Successfully connected to the server, enter the command "tensorboard --logdir=runs", the path after "=" can be modified according to your own file path, and then enter "http://127.0.0.1:12345" in the local browser , you can successfully see the loss output of model training and other information
