Rethinking Depthwise Separable Convolutions: How Intra-Kernel Correlations Lead to Improved MobileNets
Open paper
Benchmark
Model
Rank
Results
fine-grained-image-classification-on-stanford-dogs
EfficientNet-B0 (BSConv-S)
#16
Accuracy: 61.2%
image-classification-on-cifar-100
MobileNetV3-large x1.0 (BSConv-U)
#134
Percentage correct: 77.7
Rank counts only results with a code link.