List your company
ResearchFormerly WizWand

Image Classification on ImageNet (Accuracy and Energy)

80.8AccuracyTransformer
69.817672.668875.5278.3712Dec 8, 2025
Updated 8mo ago

Evaluation results

MethodLinks
TransformerArchitecture=Transformer-8-512, Param (M)=29.68, TimeStep=12025.1280.838.34
SpikingResformer+DA-SSDPArchitecture=SpikingResformer-L, Param (M)=60.38, TimeStep=42025.1279.298.76
SpikingResformer+SSDP(w/o DA)Architecture=SpikingResformer-L, Param (M)=60.38, TimeStep=42025.1279.128.89
SpikingResformerArchitecture=SpikingResformer-L, Param (M)=60.38, TimeStep=42025.1278.778.76
Spike-driven TransformerArchitecture=Spike-driven Transformer-8-768, Param (M)=66.34, TimeStep=42025.1276.326.09
SpikingformerArchitecture=Spikingformer-8-768, Param (M)=66.34, TimeStep=42025.1275.8513.68
SpikformerArchitecture=Spikformer-8-768, Param (M)=66.34, TimeStep=42025.1274.8121.477
SpikingformerArchitecture=Spikingformer-8-512, Param (M)=29.68, TimeStep=42025.1274.797.46
Spiking ResNetArchitecture=ResNet-50, Param (M)=25.56, TimeStep=3502025.1272.7570.934
SpikformerArchitecture=Spikformer-8-384, Param (M)=16.81, TimeStep=42025.1270.247.734