A keras implementation of Peddinti's paper "A time delay neural network architecture for efficient modeling of long temporal contexts". [Work in progress]
Add TDNN_layer.py to your project and include TDNN_layer. Then, use TDNN_layer as a Keras layer.
See demo.py for usage example. The file demo.py implements the network described in Peddiniti's paper.
The file demo.py implements the network described in Peddiniti's paper "A time delay neural network architecture for efficient modeling of long temporal contexts". I trained the neural network on libri speetch dataset. [Still in progress]