List your company
ResearchFormerly WizWand

Image Classification on CIFAR10 standard (test) (Top-1 Accuracy)

97.4Top-1 AccuracySpikingResformer
90.629692.387394.14595.9027Mar 21, 2024Jul 11, 2024Nov 1, 2024Feb 22, 2025Jun 14, 2025Oct 5, 2025Jan 26, 2026
Updated 8mo ago

Evaluation results

MethodLinks
SpikingResformerType=Transfer Learning, Architecture=SpikingResformer-S, #Param (M)=17.25, T=42024.0397.4
SpikformerType=Transfer Learning, Architecture=Spikformer-8-512, #Param (M)=29.08, T=42024.0397.03
SpikingResformerType=Transfer Learning, Architecture=SpikingResformer-Ti, #Param (M)=10.76, T=42024.0397.02
SpikformerType=Transfer Learning, Architecture=Spikformer-8-384, #Param (M)=16.36, T=42024.0396.64
SpikingResformerType=Direct Training, Architecture=SpikingResformer-Ti*, #Param (M)=10.79, T=42024.0396.24
TEFormerArchitecture=H-Spikformer, Step=4, SNN=true2026.0196.24
QKFormerArchitecture=H-Spikformer, Step=4, SNN=true, Evaluation=Reproduced on STEP framework2026.0195.91
SDTArchitecture=SD-Transformer, Step=4, SNN=true, Evaluation=Reproduced on STEP framework2026.0195.78
SpikingformerType=Direct Training, Architecture=Spikingformer-4-384, #Param (M)=9.28, T=42024.0395.61
Spike-driven TransformerType=Direct Training, Architecture=Spike-driven Transformer-2-512, #Param (M)=10.21, T=42024.0395.6
SpikformerType=Transfer Learning, Architecture=Spikformer-4-384, #Param (M)=9.28, T=42024.0395.54
SpikingformerArchitecture=Spikingformer, Step=4, SNN=true, Evaluation=Unified STEP framework2026.0195.53
DSRArchitecture=ResNet, Step=20, SNN=true2026.0195.4
Spiking WaveletArchitecture=SWFormer, Step=4, SNN=true, Evaluation=Unified STEP framework2026.0195.31
SpikingformerType=Direct Training, Architecture=Spikingformer-2-384, #Param (M)=5.74, T=42024.0395.22
SpikformerType=Direct Training, Architecture=Spikformer-4-384, #Param (M)=9.28, T=42024.0395.19
SpikformerArchitecture=Spikformer, Step=4, SNN=true, Evaluation=Reproduced on STEP framework2026.0195.09
Spikformer+SEMMArchitecture=Spikformer, Step=4, SNN=true, Evaluation=Unified STEP framework2026.0194.98
ResNet-19Architecture=ResNet, Step=1, SNN=false2026.0194.97
SpikformerType=Direct Training, Architecture=Spikformer-2-384, #Param (M)=5.74, T=42024.0394.8
SpikingformerType=Direct Training, Architecture=Spikingformer-4-256, #Param (M)=4.13, T=42024.0394.77
DspikeType=Direct Training, Architecture=ResNet-18, #Param (M)=11.21, T=62024.0394.25
TIMArchitecture=Spikformer, Step=4, SNN=true, Evaluation=Reproduced on STEP framework2026.0194.2
SpikformerType=Direct Training, Architecture=Spikformer-4-256, #Param (M)=4.13, T=42024.0393.94
DspikeType=Direct Training, Architecture=ResNet-18, #Param (M)=11.21, T=42024.0393.66
SNASNetArchitecture=ConvNet, Step=5, SNN=true2026.0193.64
PLIFType=Direct Training, Architecture=6 Conv, 2 FC, #Param (M)=36.71, T=82024.0393.5
PLIFArchitecture=ConvNet, Step=8, SNN=false2026.0193.5
STBP-tdBNType=Direct Training, Architecture=ResNet-19, #Param (M)=12.54, T=62024.0393.16
DspikeType=Direct Training, Architecture=ResNet-18, #Param (M)=11.21, T=22024.0393.13
STBP-tdBNType=Direct Training, Architecture=ResNet-19, #Param (M)=12.54, T=42024.0392.92
tdBNArchitecture=ResNet, Step=4, SNN=true2026.0192.92
DIET-SNNArchitecture=VGGNet, Step=5, SNN=true2026.0192.7
STBP-tdBNType=Direct Training, Architecture=ResNet-19, #Param (M)=12.54, T=22024.0392.34
ViTArchitecture=ViT, Step=1, SNN=false, Evaluation=Unified STEP framework2026.0190.89