rithulkamesh/music_classifier

AI Summary: Experiment with different CNN architectures (ResNet, VGG, MobileNet, EfficientNet), CNN-RNN hybrids, and Transformer-based models (AST, ViT) to improve the accuracy and efficiency of a music genre classification model. This involves implementing or importing the chosen architectures, training them on the FMA Medium dataset, comparing their performance (accuracy, F1 score, inference time) against a baseline, and selecting the best-performing model for integration into the existing Streamlit application. Optional exploration of self-supervised pretraining is also suggested.