Deep CNN model training

Can anyone comment on this model learning,
Is it overfitting, I am already using BN, Dropout.
Or should I simply reduce my epochs to approx 50, is it fine.
Actually, I want to know what else I can do to improve the performance.


Yes, a perfect training accuracy and a gap to the validation accuracy would indicate overfitting.
Generally you could increase the regularization through dropout, a smaller model, more aggressive data augmentation etc., but as always it depends on your use case what would work best.


What is meant by regressive augmentation, is it allowed in research paper writing? Do people mention it either. Actually during my studies I found results mismatched with mine from so many paper , are they using augmentation?
I am using this for audio classification.
Can I train the model after 40 epochs coz after that nothing is changed much sir?