no code implementations • 17 Jul 2020 • Nir Ailon, Omer Leibovich, Vineet Nair
Motivated by these facts, we propose to replace a dense linear layer in any neural network by an architecture based on the butterfly network.
Decoder Representation Learning