A Benchmark of Dynamical Variational Autoencoders applied to Speech Spectrogram Modeling - 42Papers