Genrify

Overview

Genrify is a Convolutional Neural Network (CNN)-based ensemble learning model designed for the classification of three-second audio clips into one of ten genres:

• Blues
• Classical
• Country
• Disco
• Hip-hop
• Jazz
• Metal
• Pop
• Reggae
• Rock

Model Architecture

Our model employs ensemble learning, where four parallel CNNs work in a majority vote scheme to make the final prediction. Each weak learner CNN adopts a different preprocessing approach for the three-second WAV file input. The four preprocessing methods used are:

• Spectrogram
• Mel-spectrogram
• Chroma
• MFCC

Resources

For more detailed information about the project, please refer to our presentation:

• Genrify Presentation

For our comprehensive final report, you can access it through this link:

• Genrify Final Report

Feel free to explore these resources for a more in-depth understanding of the Genrify project.

---