search

keunwoochoi / music-auto_tagging-keras

Music auto-tagging models and trained weights in keras/theano
MIT License
616 stars 142 forks source link

sound file duration #38

Open zzstefan opened 5 years ago

zzstefan commented 5 years ago

Hi, If I want to extract a music file with duration of 2s. Can I use compact_cnn to extract the music sound features? I see the duration in your code is 29.s.

keunwoochoi commented 5 years ago

In short, no, but probably you can use https://github.com/keunwoochoi/transfer_learning_music, especially https://github.com/keunwoochoi/transfer_learning_music/tree/master/keras2_model this one. Check out the ipython notebook file, where the model is defined with varying input temporal length.