On the role of synaptic stochasticity in training low-precision neural networks

Open in new window