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.