How to detect overfitting in neural network?
10 views (last 30 days)
Chunru on 28 Apr 2023
It looks that the overfitting is not occuring in your case since the validation performance still keep improving.
Overfitting occurs when training performance is still improving while validation performance becomes worse.
For your case, both training and validation performances are still improving (error reducing) so you can keep training the model for further performance improvement.
The "best" (circle on green curve) shows that the latest model is the best. Thus there is still room for further improvement.