Skip to content
ml-recall

#gru

4 questions from Recurrent Neural Networks

Timed test

4 questions · 4 min · score and explanations at the end

Take a test
Q 1 / 4Score 0 / 0

Question 1: Short answer

mediumAI-checkedRecurrent Neural Networks

For the same input size dd and hidden size hh, every gate or candidate state of a recurrent layer costs one block XtWx⋅+Ht−1Wh⋅+b⋅\mathbf{X}_t\mathbf{W}_{\textrm{x}\cdot} + \mathbf{H}_{t-1}\mathbf{W}_{\textrm{h}\cdot} + \mathbf{b}_\cdot of matrix products. Per time step, a GRU layer needs about times and an LSTM layer about times the multiply-adds of a vanilla RNN layer.

Fill in every blank (hints in the boxes). Enter moves to the next blank.