Val loss nan keeps nan

[ZYICE111]

Thank you for your excellent work. I tried to apply your work to our own data to achieve binary classification, but the val loss calculation kept nan all the time, resulting in no change in accuracy. Do you know what the problem is? Looking forward to your reply. Thank you very much!

[happywt]

Hello, can you tell me how to configure your conda virtual environment? Is it enough to follow the environment. yml file provided by the author?