CutMix (DeepLab v2, ImageNet pre-trained)

Reported on 2 benchmarks across 2 tasks · 1 paper · 2 SOTA

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

Medical1 result

Semantic SegmentationonCityscapes 100 samples labeled
Validation mIoU· 2019-06-05
51.2
best: 76.2 (SemiVL (ViT-B/16))
SOTA
Semi-supervised semantic segmentation needs strong, varied perturbations arXiv:1906.01916

Audio1 result

10-shot image generationonCityscapes 100 samples labeled
Validation mIoU· 2019-06-05
51.2
best: 76.2 (SemiVL (ViT-B/16))
SOTA
Semi-supervised semantic segmentation needs strong, varied perturbations arXiv:1906.01916