This is a small project me and my friend did for a deep learning assignment. This project focuses on trying different models for the best accuracy we can get at the time. The feature extraction of audio data is side challenge that we decided to take on.
The test set is heavily skewed to the male side (92.6%) so the final accuracy being at 96.43% is a slight improvement. If possible, I'd love to see the result better in a less-skewed dataset and maybe test the model out in an application environment.