Skip to content

Code relating to paper submitted for review at INTERSPEECH 2018

Notifications You must be signed in to change notification settings

ZackHodari/IS18_control_space

Repository files navigation

IS18_control_space

Code relating to paper submitted for review at INTERSPEECH 2018

Makes use of modNN, a TensorFlow interface that allows for input, output, and model handlers to be combined together as modules.


Models are trained using run.py using the run_task function which takes a config (python dictionary) as input.

Example config included at setup.py, two formats possible;

  • Sequential computational graph (SimpleModel): one input, one output, sequential NN modules
  • Customisable computational graph (GraphModel): requires handlers to be given names and an adjacency list to be defined in the config.

About

Code relating to paper submitted for review at INTERSPEECH 2018

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published