List your company
ResearchFormerly WizWand

Image Classification on ImageNet-1K Regular (val)

0.845Top-1 AccuracySwin-B
0.76180.78340.8050.8266Mar 25, 2021
Updated 8mo ago

Evaluation results

MethodLinks
Swin-Bimage size=384x384, #param.=88M, FLOPs=47.0G, throughput (image/s)=84.72021.030.845
EffNet-B7image size=600x600, #param.=66M, FLOPs=37.0G, throughput (image/s)=55.12021.030.843
EffNet-B6image size=528x528, #param.=43M, FLOPs=19.0G, throughput (image/s)=96.92021.030.84
EffNet-B5image size=456x456, #param.=30M, FLOPs=9.9G, throughput (image/s)=169.12021.030.836
Swin-Bimage size=224x224, #param.=88M, FLOPs=15.4G, throughput (image/s)=278.12021.030.835
DeiT-Bimage size=384x384, #param.=86M, FLOPs=55.4G, throughput (image/s)=85.92021.030.831
Swin-Simage size=224x224, #param.=50M, FLOPs=8.7G, throughput (image/s)=436.92021.030.83
RegNetY-16Gimage size=224x224, #param.=84M, FLOPs=16.0G, throughput (image/s)=334.72021.030.829
EffNet-B4image size=380x380, #param.=19M, FLOPs=4.2G, throughput (image/s)=349.42021.030.829
DeiT-Bimage size=224x224, #param.=86M, FLOPs=17.5G, throughput (image/s)=292.32021.030.818
RegNetY-8Gimage size=224x224, #param.=39M, FLOPs=8.0G, throughput (image/s)=591.62021.030.817
EffNet-B3image size=300x300, #param.=12M, FLOPs=1.8G, throughput (image/s)=732.12021.030.816
Swin-Timage size=224x224, #param.=29M, FLOPs=4.5G, throughput (image/s)=755.22021.030.813
RegNetY-4Gimage size=224x224, #param.=21M, FLOPs=4.0G, throughput (image/s)=1156.72021.030.8
DeiT-Simage size=224x224, #param.=22M, FLOPs=4.6G, throughput (image/s)=940.42021.030.798
ViT-B/16image size=384x384, #param.=86M, FLOPs=55.4G, throughput (image/s)=85.92021.030.779
ViT-L/16image size=384x384, #param.=307M, FLOPs=190.7G, throughput (image/s)=27.32021.030.765