ResNet-34 Implementation

I implemented a ResNet architecture from scratch, gaining a deeper understanding of the inner processes of neural networks. The structure was used to classify images of cats with different fur patterns, and reached a testing accuracy of up to 80% with 8 classes. This was a part of the UNSW Course Neural Networks (COMP9444) for which I got a High Distinction grade.

Experiences Gained: Data Augmentation, Knowledge of Optimisers, Setting up a Pytorch Project

Oltan Sevinc