iPython notebooks for feature extraction of melspectrograms and training (via transfer learning) of audio event classifier on ESC-50 dataset. Base model used in VGG 16. In order to run, ESC-50 dataset must be downloaded locally and path to meta/esc50.csv needs to be specified.
lvaleriu / cnn-audio-classifier-with-keras-tensorflow Goto Github PK
View Code? Open in Web Editor NEWThis project forked from adanrivas/cnn-audio-classifier-with-keras-tensorflow
ipython notebooks for feature extraction and training of audio event classifier on ESC-50 dataset.