Training spiking multi-layer networks with surrogate gradients on an analog neuromorphic substrate

Open in new window