The Hidden Markov Model Toolkit (HTK)
The Hidden Markov Model Toolkit (HTK) is a portable toolkit for building and manipulating hidden Markov models. HTK is primarily used for speech recognition research although it has been used for numerous other applications including research into speech synthesis, character recognition and DNA sequencing. HTK is in use at hundreds of sites worldwide.
The HTK Book for HTK Version 3.5 has DNN-HMM Systems and the chapter 14 dedicated to Artificial Neural Networks.
A DNN-HMM system A deep neural network (DNN) is a multi-layer perceptron (MLP) with many hidden layers.
CUED-RNNLM toolkit present an implementation of recurrent neural network l anguage models (RNNLMs) training (on GPU), and efficient evaluation(on CPU). CUED-RNNLM toolkit present an implementation of RNNLM training (on GPU), and efficient evaluation(on CPU). http://mi.eng.cam.ac.uk/projects/cued-rnnlm/
- organise installation htk3.4
- add examples for htk3.5