Transfer Learning with Binary Neural Networks

Sam Leroux, Steven Bohez, Tim Verbelen +3

Previous work has shown that it is possible to train deep neural networks with low precision weights and activations. In the extreme case it is even possible to constrain the netwo…