Base Code (Version 1)
We created below model to begin with and applied it on 101 images, 76 in training and 25 in
validation.
Used epochs = 20, training steps per epoch = 25 and 30 in validation steps and got accuracy of 61-62%
and validation accuracy of 36%.