wandb connection problem

[zhaoyongchao2019]

Hello, last week I was able to run this program normally, but this week I don't know what went wrong. It keeps throwing an error and cannot access api.wandb.ai. I want to try setting wandb to offline or simply not using wandb, but the information I found online couldn't solve my problem. I hope you can tell me which part of the code I should modify. Thank you very much.

[LinWeizheDragon]

You can first try connecting to wandb with some toy codes to rule out the possibility of local network connection issues And then you can set WANDB_MODE to offline to disable online logging. https://docs.wandb.ai/guides/track/environment-variables

[zhaoyongchao2019]

Thank you, the error disappeared inexplicably today, and there is no problem now.

[zhaoyongchao2019]

你好，我又遇到这个问题了，我在main.py开头添加了os.environ["WANDB_MODE"] = "offline"，但依旧会报这个错误，请问是我添加的位置有错吗？如果你知道原因请告诉我

[LinWeizheDragon]

I did not encounter this problem recently. It may be that the connectivity in China is too unstable. If you want to enforce offline mode, you can do by running wandb offline first under the working directory, and then run your experiments - there might be some codes that need to be altered to work with offline mode but it should be fine. After you finish the experiment, you can do wandb online and then wandb sync --sync-all to upload these runs when the connection restores.