CutMix

Reported on 5 benchmarks across 3 tasks · 2 papers · 4 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical2 results

Semantic SegmentationonADE20K 1/16 labeled
Validation mIoU· 2019-06-05
29.8
best: 46.7 (UniMatch V2)
SOTA
Semi-supervised semantic segmentation needs strong, varied perturbations arXiv:1906.01916
Semantic SegmentationonADE20K 1/32 labeled
Validation mIoU· 2019-06-05
26.2
best: 45 (UniMatch V2)
SOTA
Semi-supervised semantic segmentation needs strong, varied perturbations arXiv:1906.01916

Audio2 results

10-shot image generationonADE20K 1/16 labeled
Validation mIoU· 2019-06-05
29.8
best: 46.7 (UniMatch V2)
SOTA
Semi-supervised semantic segmentation needs strong, varied perturbations arXiv:1906.01916
10-shot image generationonADE20K 1/32 labeled
Validation mIoU· 2019-06-05
26.2
best: 45 (UniMatch V2)
SOTA
Semi-supervised semantic segmentation needs strong, varied perturbations arXiv:1906.01916

Computer Vision1 result

Image ClassificationonOmniBenchmark
Average Top-1 Accuracy· 2019-05-13
31.1
best: 47.6 (NOAH-ViTB/16)
CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features arXiv:1905.04899