Scaling Spike-driven Transformer with Efficient Spike Firing Approximation Training

TL;DR

This paper introduces a Spike Firing Approximation method and an efficient spike-driven Transformer architecture, significantly improving the performance and efficiency of Spiking Neural Networks on large-scale visual tasks, making them competitive with traditional neural networks.

Contribution

The work proposes a novel Spike Firing Approximation technique and a spike-driven Transformer design, enabling high-performance, low-power SNNs that scale effectively and match ANN accuracy.

Findings

Achieved state-of-the-art top-1 accuracy of 86.2% on ImageNet-1k with 173M parameters.

Outperformed existing SNNs by 7.2% on ImageNet-1k.

Improved training time and inference energy efficiency by 4.5× and 3.9×, respectively.

Abstract

The ambition of brain-inspired Spiking Neural Networks (SNNs) is to become a low-power alternative to traditional Artificial Neural Networks (ANNs). This work addresses two major challenges in realizing this vision: the performance gap between SNNs and ANNs, and the high training costs of SNNs. We identify intrinsic flaws in spiking neurons caused by binary firing mechanisms and propose a Spike Firing Approximation (SFA) method using integer training and spike-driven inference. This optimizes the spike firing pattern of spiking neurons, enhancing efficient training, reducing power consumption, improving performance, enabling easier scaling, and better utilizing neuromorphic chips. We also develop an efficient spike-driven Transformer architecture and a spike-masked autoencoder to prevent performance degradation during SNN scaling. On ImageNet-1k, we achieve state-of-the-art top-1 accuracy of 78.5\%, 79.8\%, 84.0\%, and 86.2\% with models containing 10M, 19M, 83M, and 173M parameters, respectively. For instance, the 10M model outperforms the best existing SNN by 7.2\% on ImageNet, with training time acceleration and inference energy efficiency improved by 4.5$\times$ and 3.9$\times$, respectively. We validate the effectiveness and efficiency of the proposed method across various tasks, including object detection, semantic segmentation, and neuromorphic vision tasks. This work enables SNNs to match ANN performance while maintaining the low-power advantage, marking a significant step towards SNNs as a general visual backbone. Code is available at https://github.com/BICLab/Spike-Driven-Transformer-V3.