I am trying to predict remaining useful life (RUL) from temporal data with multilayered LSTM and obtaining the following curve:

enter image description here

Looks like after first several epochs performance stops to improve and starts oscillating. Is it known situation?


Two options are most likely:

  • Your learning rate is too high to further improve. Try decreasing it by 2-10 times and see what happens. Potentially using a learning rate scheduler is beneficial - start with high learning rate but decrease it after a few epochs.
  • You don't have enough capacity in the network or you are not inputting things correctly. Hard to say without knowing more details and seeing the code

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.