CNN with data augmentation from Cinic10
This CNN was a prooject and consist of training a CNN simple and with Data augmentation.
It was thought to show Occam's Principle (you'll see how and why in the code).
It uses Cifar 10 and a set of subsets that I (manually created from the whole set !!! :D...) from Cinic 10 to test results in data augmentation. Those are available to download.