Among its features you can find:
* A lot of things in gradient machines, that is, machines which could be learned with a gradient descent. This includes multi-layered perceptrons, radial bas is functions, mixtures of experts, convolutional networks and even time-delay neural networks. In fact a lot of "modules" are available that you can plug as you want to get what you need.
* Support vector machines, in classification and regression. As fast as the old stand-alone program SVMTorch II, but with the powerful environment of the library.
* Ensemble models such as bagging or adaboost.
* Non-parametric models such as K-nearest-neighbors, Parzen regression and Parzen density estimator.
* Distributions stuff, like Kmeans, Gaussian mixture models, hidden Markov models, input-output hidden Markov models, and Bayes classifier.
* Speech recognition tools (Embedded training and large vocabulary decoding).