WaveMix: A Resource-efficient Neural Network for Image Analysis

Benchmark Model Rank Results
image-classification-on-caltech-256WaveMixLite-256/7#4Accuracy: 54.62
image-classification-on-cifar-10WaveMixLite-144/7#88Percentage correct: 97.29
image-classification-on-cifar-100WaveMixLite-256/7#65Percentage correct: 85.09
image-classification-on-cifar-100WaveMix-Lite-256/7#163Percentage correct: 70.20
image-classification-on-emnist-balancedWaveMixLite-128/7#3Accuracy: 91.06
image-classification-on-emnist-digitsWaveMixLite-112/16#2Accuracy (%): 99.82
image-classification-on-emnist-lettersWaveMixLite-112/16#1Accuracy: 95.96
image-classification-on-fashion-mnistWaveMixLite#6Percentage error: 5.68
image-classification-on-imagenetWaveMix-192/16 (level 3)#902Top 1 Accuracy: 74.93%
image-classification-on-places365-standardWaveMix-240/12 (level 4)#4Top 1 Accuracy: 56.45
image-classification-on-stl-10WaveMixLite-256/7#59Percentage correct: 70.88
image-classification-on-svhnWaveMixLite-144/15#5Percentage error: 1.27
image-classification-on-tiny-imagenet-1WaveMixLite-144/7#10Validation Acc: 77.47%
semantic-segmentation-on-cityscapes-valWaveMix#30mIoU: 82.7
semantic-segmentation-on-cityscapes-valWaveMix-256/16 (Level-4)#32mIoU: 82.60