While training (my 3d_conv NN), the accuracy of the train set is high - 65% (don't mind the validation accuracy yet):

enter image description here

But - when I predict the labels of the TRAIN set on the trained model, I get a terrible accuracy (instead of the anticipated over-fitted result of 65%):

enter image description here

How is that possible? Why would it happen?

Another question would be on the validation set - why doesn't it improve at all?

The data is split randomly between the train and validation set using sklearn's train_test_split() with the "stratify" option.

Any help would be much appreciated.

New contributor
AlonBA is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
  • 1
    $\begingroup$ Your model only predict no movement... there is a problem somewhere in your training or inference phase. $\endgroup$
    – lcrmorin
    Sep 21 at 16:36


Your Answer

AlonBA is a new contributor. Be nice, and check out our Code of Conduct.

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy