I am training an LSTM for different series (passing them all at once).
While training I print the MAE of each training batch and validation batch. At the end of the epoch, I print the "epoch error" as the mean of the errors in all the batches (in the original level of the variable). Let's say that this number is around 2.
Then I run the model over the validation set and I get a much higher MAE, like 10.
How is possible that the mean of the batches' MAE of validation set is much smaller than the MAE of the whole validation dataset? What can be causing this? Thanks.