I have a dataset of video sequences, I have trained them, and calculated the training loss using mean square error, but my training loss is decreasing down very fast. Like 0.06-0.02. Is it just fine or not?

Here's my graph representing both training and validation loss


  • $\begingroup$ This means it is very easy for your model to learn the task. But only your validation set loss will tell whether it is overfitting or not. $\endgroup$ Jun 4, 2021 at 7:48
  • $\begingroup$ I have edited my question. kindly review $\endgroup$
    – TariqS
    Jun 4, 2021 at 8:18
  • $\begingroup$ So your model is getting slightly overfit, becuase train loss is lower than the val loss. You can look into techniques to avoid overfitting. And at the end, validation set is your target. If your model is doing satisfactory performance of val set, then training is successful. Also I am not clear with what so you mean by "Is it just fine or not?" $\endgroup$ Jun 4, 2021 at 8:44
  • $\begingroup$ I suggest to continue training for a higher number of epochs since, as your graph shows, val loss continues to decrease. $\endgroup$
    – Jonathan
    Jun 4, 2021 at 10:34
  • $\begingroup$ So you are saying that I should stop when val_loss starts to converge. Is it? $\endgroup$
    – TariqS
    Jun 4, 2021 at 13:36


Your Answer

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