fraunhofer-iais/dense-hmm

DenseHMM is a modification of Hidden Markov Models (HMMs) that allows to learn dense vector representations of both the hidden states and the observables via gradient-descent methods.

PythonMIT