I recently found an interesting paper on what it really means for a recurrent neural network (RNN) to be deep here. Depth can be added in several different ways (state to state, input to state, etc.) and the authors give several interesting architectures. I've been looking (via Google) for packages that support creating these kinds of structures in an intuitive way that is quick/easy to implement but I have come up empty thus far. Does anyone know of a freely available package that supports the various structures detailed in the reference paper?