histocartography / hact-net

MIT License
70 stars 14 forks source link

Val loss nan keeps nan #12

Open ZYICE111 opened 10 months ago

ZYICE111 commented 10 months ago

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! 003a79696733ba37a6662141be10993

happywt commented 8 months ago

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?