There is a newer version of the record available.

Published January 17, 2018 | Version 0.1.b2
Software Open

NickleDave/hybrid-vocal-classifier 0.1.b2

Authors/Creators

Description

  • can now train scikit-learn models with 'predict probability' set to True
  • flatwindow model in Keras works
  • svm feature extraction deals appropriately with segments from which it can't extract features by returning zeros instead; this reproduces approach of Tachibana Okinaya paper

Files

NickleDave/hybrid-vocal-classifier-0.1.b2.zip

Files (187.4 MB)

Name Size Download all
md5:aece6f150793422ec8e21b9ebcdc8dee
187.4 MB Preview Download

Additional details

Related works

Is supplement to
https://github.com/NickleDave/hybrid-vocal-classifier/tree/0.1.b2 (URL)