Determining the loss function according to the problem

Question

I\'m trying to solve a problem using LSTM. I\'m striving to code the LSTM class myself, I don\'t use Keras, etc. frameworks.

I need to set the loss function but I\'m a li