This is the setup for the Dynamic Time Warping (DTW) system developed at the Spoken Language Systems Lab. You can fork the code in Github [1] or directly download the zip [2].
To run the basic example, you will also need the posterior features available here [3].