Here we classify various species of birds, using PyTorch.
The dataset we used contains 510 different species of birds, and we are required to classify them accordingly. We didn't use the entire dataset given, rather used 10 different species, i.e, we have 10 classes to classify (to make it easier on the machine). We built a CNN (Convolutional Neural Network) from scratch, and trained it using the train images, and achieved a 96% accuracy on the test set.
The subset of the entire dataset is available here:-
Training Set
Test Set