Different losses while training

[jomiguelcarv]

Im trainning on a custom dataset and having some difficulty understanding how to properly track the trainning process. Could someone please elaborate a little bit on the differences between loss, val/rec_loss_step, val/aeloss_step, train/aeloss_epoch, etc? Which ones should i be looking at with most attention?

[nicolasfischoeder]

well, aelosses are the encoders losses i assume, one for each step and another one for a whole epoch. however i do not know what rec stands for. Reconstruction maybe?