SGS-SNN

Surrogate Gradient Scaling for Directly Training Spiking Neural Networks

This paper has been submitted to Applied Intelligence. The complete code will be made public when our paper is accepted.

Figures

Requirements

• Python 3.9.7
• Torch 1.10.1
• Torchvision 0.11.2
• Numpy 1.22.0

Datasets

• CIFAR10
• CIFAR100
• DVS-CIFAR10

TO DO

• layers.py: surrogate gradient scaling function corresponds to Eq. (5) in our paper.
• my_main.py: Specify the dataset location ; Specify .pth file location. Inference model
• train.py: release soon

Usage

• run my_main.py on CIFAR10, T=2, can achieve accuracy of 94% where in our ablation study of our manuscript.

python my_main.py