0
$\begingroup$

Is this a good result? How do you print the model accuracy (in %) from this graph?

enter image description here

$\endgroup$
6
  • $\begingroup$ You need to know what your loss function is $\endgroup$
    – Allan
    Jul 29 at 13:11
  • $\begingroup$ why is your training data results moving inverse to the test data training results $\endgroup$ Jul 29 at 16:47
  • $\begingroup$ @Allan how to do that? $\endgroup$
    – X002
    Jul 31 at 5:35
  • $\begingroup$ @GoldenLion any suggestions? $\endgroup$
    – X002
    Jul 31 at 5:36
  • $\begingroup$ try to identify the outlier data using standard deviation $\endgroup$ Jul 31 at 14:19

1 Answer 1

2
$\begingroup$

That is the graph of the loss curves. It clearly shows that the model is overfitting because the test curve is going up. This is bad. It means that your model will perform badly in data that is not very very similar to the training data.

This graph is not directly related to the accuracy. You cannot know the accuracy by looking at the loss.

$\endgroup$
3
  • $\begingroup$ @neo Thanks for the prompt reply. How to fix it? $\endgroup$
    – X002
    Jul 31 at 5:33
  • $\begingroup$ Some things you may try are reducing your model capacity (e.g. reducing the embedding size) or applying regularization (e.g. dropout). $\endgroup$
    – noe
    Jul 31 at 7:27
  • $\begingroup$ Also, please consider upvoting and marking the answer as correct. $\endgroup$
    – noe
    Jul 31 at 7:28

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.