Table 2.
Accuracy of spiking neural network (SNN) trained using BlocTrain and end-to-end spike-based backpropagation through time (BPTT) methods on the CIFAR-100 dataset.
| Model | Training method | Dataset size | %Accuracy |
|---|---|---|---|
| SNN w/ 8 layers (Thiele et al., 2020) | End-to-end ANN-based SpikeGrad | 50,000 | 64.69 |
| VGG-11 (Rathi et al., 2020) | ANN-SNN and end-to-end STDB | 50,000 | 67.87 |
| ResNet-50 (Ledinauskas et al., 2020) | End-to-end Spike BP | 50,000 | 58.5 |
| ResNet-11 (our work) | BlocTrain | 40,000 | 58.21 |
| ResNet-11 (our work) | BlocTrain-base | 50,000 | 62.03 |
| VGG-16 (our work) | BlocTrain | 50,000 | 61.65 |
The bold values are used to highlight the results reported in this work over prior works.