I am trying to build a classification algorithm having 28 classes. These classes consists of Logo of companies like adidas , Nike etc. I have very low dataset below than 100 images and greater than 70 images. I have trained CNN model but not got decent results. We all know that if error of the test data is high , then model is overfitted and if model performs poorly on test set it is also overfitted. If training loss decreases , and then starts increasing gradually model is also overfitted. There are any scenarios to identify the problem . Which identification should I choose in my case.
$\begingroup$
$\endgroup$
1
-
$\begingroup$ Do you mean by "identification"? What are your thoughts? Can you make this useful for others who aren't working on the exact same dataset you are? $\endgroup$– D.W. ♦Jan 17 at 20:34
Add a comment
|