modelling sequence of a sequence

[issafuad]

encoder for a sequence of sentences
I am wondering if there is a way to have the sequence encoder taking a sequence of sentences (word sequence / or just a vector that is already generated for that sentence ) where there is a sentence at each time step. this is with the final aim to have the sequence combiner combining multiple sequence features where some are a sequence of sentences and others are a sequence of numbers, categories etc.