-
sstdp_module (
/code/sstdp_module.py
)- SpikeOnceNeuron: The neuron that will only spike once for each sample. Such neuron reduces the spike density while maintain spike information
- stdp_update: The STDP update rule that direct the stdp update with gradient.
- stdp_linear_container: The linear spiking neuron layer container.
- StdpLinear: The callable linear model.
- stdp_conv2d_container: The 2d convolution spiking neuron layer container.
- StdpConv2d: The callable 2d convolution model.
-
sstdp_train (
/code/sstdp_train.py
) run single, with parameterspython sstdp_train.py --threshold 100 --result_dir test_train/ --weight_decay 1e-5 --learning_rate 10
We now have a paper, titled "SSTDP: Supervised Spike Timing Dependent Plasticity for Efficient Spiking Neural Network Training", which is published in Frontiers in Neuroscience, Section Neuromorphic Engineering.
@article{liu2021sstdp,
title={SSTDP: Supervised Spike Timing Dependent Plasticity for Efficient Spiking Neural Network Training},
author={Liu, Fangxin and Zhao, Wenbo and Chen, Yongbiao and Wang, Zongwu and Yang, Tao and Li, JIANG},
journal={Frontiers in neuroscience},
year={2021},
pages={1413},
publisher={Frontiers}
}
- Coming soon: Updated Code.