Shape of test / validation data for RNN / LSTM

Question

I'm training a RNN using LSTM, I understand how to present the training data as a 3D tensor in the form [batch size, timesteps, features], but I'd like some clarification for the test/validation data; am I right in thinking the test/validation data is also shaped in the same way as a training batch that contains arrays of timesteps and features / expected output?

Answer 1

To answer my own question: yes, I form my validation data in the same way as training data.