How to increase my CNN model accuracy from 72% to 90%

Hi ,
I have made a CNN model.3 clases total in dataset
Traiining data include : 1500 training images ( ,500 images from each class) and
testing 600 testing images ( 200 imags from each class)
epochs = 10
What should be datasized and epochs setting for best accuracy.Please help

It’s hard to predict the best hyperparameters as they will depend on your particular problem.
What have you tried so far and where did you get stuck?

Thankyou for replying ,the accuracy has improved .

Well, we like to hear your story, why the accuracy is improved, tell us about that…

this worked : increased the training images data size from 500 to 700 in each class ,so the total training dataset size increased and complex .