LaSNN: 深度脉冲神经网络中的有效高效培训之层级 ANN 到 SNN 蒸馏（英文版）

英文标题：LaSNN: Layer-wise ANN-to-SNN Distillation for Effective and Efficient Training in Deep Spiking Neural Networks中文摘要：通过引入 attention 机制和采用分层知识蒸馏的方法，提出了新的 SNN 训练框架 LaSNN，实现了与 ANN 比较的竞争性准确率和比转换的 SNN 快 20 倍的推理速度，并且具有灵活性和可扩展性。英文摘要：Spiking Neural Networks (SNNs) are biologically realistic and practicallypromising in low-power computation because of their event-driven mechanism.Usually, the training of SNNs suffers accuracy loss on various tasks, yieldingan inferior performance compared with ANNs.