no code implementations • 29 Sep 2021 • Clément Metz, Thibault Allenet, Johannes Christian Thiele, Antoine Dupret, Olivier Bichler
In particular, we achieve ImageNet classification results close to full precision on the popular Resnet-18/50, with only 0. 5% and 5% accuracy drop for the 4-bit weights and 3-bit weights model architectures respectively.
no code implementations • ICLR 2020 • Johannes Christian Thiele, Olivier Bichler, Antoine Dupret
In this work, we analyze how the gradient can be discretized into spike events when training a spiking neural network.
no code implementations • 11 Mar 2019 • Johannes C. Thiele, Olivier Bichler, Antoine Dupret, Sergio Solinas, Giacomo Indiveri
Our architecture is the first spiking neural network architecture with on-chip learning capabilities, which is able to perform relational inference on complex visual stimuli.