This paper introduces a post-training encoding method that reduces the number of time-steps required for spiking neural networks.