EfficientNetV2: Smaller Models and Faster Training

Open paper
Benchmark Model Rank Results
image-classification-on-stanford-carsEfficientNetV2-L#2Accuracy: 95.1
image-classification-on-stanford-carsEfficientNetV2-M#3Accuracy: 94.6
image-classification-on-stanford-carsEfficientNetV2-S#6Accuracy: 93.8
image-classification-on-cifar-100EfficientNetV2-L#18Percentage correct: 92.3
image-classification-on-cifar-100EfficientNetV2-M#19Percentage correct: 92.2
image-classification-on-cifar-100EfficientNetV2-S#28Percentage correct: 91.5
image-classification-on-cifar-10EfficientNetV2-L#19Percentage correct: 99.1
image-classification-on-cifar-10EfficientNetV2-M#30Percentage correct: 99.0
image-classification-on-cifar-10EfficientNetV2-S#43Percentage correct: 98.7
image-classification-on-flowers-102EfficientNetV2-L#24Accuracy: 98.8
image-classification-on-flowers-102EfficientNetV2-M#27Accuracy: 98.5
image-classification-on-flowers-102EfficientNetV2-S#36Accuracy: 97.9
image-classification-on-imagenetEfficientNetV2-XL (21k)#104Top 1 Accuracy: 87.3%Number of params: 208MGFLOPs: 94
image-classification-on-imagenetEfficientNetV2-L (21k)#129Top 1 Accuracy: 86.8%Number of params: 120MGFLOPs: 53
image-classification-on-imagenetEfficientNetV2-M (21k)#177Top 1 Accuracy: 86.2%Number of params: 54MGFLOPs: 24
image-classification-on-imagenetEfficientNetV2-L#218Top 1 Accuracy: 85.7%GFLOPs: 53
image-classification-on-imagenetEfficientNetV2-M#274Top 1 Accuracy: 85.1%
image-classification-on-imagenetEfficientNetV2-S (21k)#294Top 1 Accuracy: 84.9%Number of params: 22MGFLOPs: 8.8
image-classification-on-imagenetEfficientNetV2-S#403Top 1 Accuracy: 83.9%

Rank counts only results with a code link.