0

I would like to be sure of whether the model is overfitting or undercutting. Being new to this, is there any specific point to identify when to stop the training process. Any help in this regard would be helpful. Thanks.enter image description here

enter image description here

Anna
  • 1
  • 2
    Welcome to SE, I have one question about your graphs : Are train and test curves mixed up ? Usually (and by usually, I mean all of the time) train curve has a lower loss and a better accuracy than test curves, which is not your case. But to give my opinion on your question, I'd say that there is no overfitting here (*overfitting is a decrease of test accuracy after too many epochs*). But it is clear as well that there your model is not improving after the 200th epochs (which is already a lot of epochs), so I would have stopped the training after 200 epochs. – Ubikuity Jun 09 '21 at 15:58
  • Thank you! one more query, would you call the above a good fit model? – Anna Jun 09 '21 at 18:36
  • 1
    You're welcome, about accuracy this all depends on your task, the easier the task the easier it is to have a good accuracy. But your final accuracy is 98%, which is definitly a great accuracy (you can consider above 95% as great accuracies). Tbh, I wish I had 98% accuracy on my models. – Ubikuity Jun 09 '21 at 18:40
  • Again, thank you so much. – Anna Jun 09 '21 at 18:43

0 Answers0