Any ways to use a RNN cell with Universal Sentence Encoder

[thomasmartin1783]

USE (Universal Sentence Encoder) takes a variable length text as input and outputs a 512 dimensional feature vector as we know.

RNN cell takes 3 parameters as input.
[batch_size, timesteps, output_size]

So, my question is after encoding with Universal Sentence Encoder how can we pass these outputs to a RNN cell with takes 3 parameters while we only get (None, 512) from USE