This is a simple image classifier that can distinguish between alligators and crocodiles. After cleaning the data and finding the optimal learning rate, the accuracy achieved was 86%.
The Convolutional Neural Network (CNN) architecture used was Resnet-34, which comes with pre-trained weights from Imagenet.
You can test it yourself in a web application I created here: https://croc-or-alli.herokuapp.com/. Just upload a picture of an alligator or a crocodile and the algorithm will give you its prediction along with its confidence value.
