MGR

Music Genre Recognition using Convolutional Neural Network.

Input :

• Dataset directory containing folders equal to the number of classes
• csv file [Path, Filename, Genre]

Output :

• saved model at given directory

Workflow :

Reads mp3/wav file -> Data Augmentation if required -> extracts mfcc -> split into training/validation/testing memory map files -> creates tf dataset based on those files -> trains the CNN model -> saves the model