Robust LR

Reported on 4 benchmarks across 1 task · 1 paper · 2 SOTA

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

Computer Vision4 results

Image Classificationonmini WebVision 1.0
Top-1 Accuracy· 2021-12-06
81.84
best: 84.16 (LRA-diffusion (CLIP ViT))
SOTA
Two Wrongs Don't Make a Right: Combating Confirmation Bias in Learning with Label Noise arXiv:2112.02960
Image Classificationonmini WebVision 1.0
Top-5 Accuracy· 2021-12-06
94.12
best: 94.84 (PSSCL (130 epochs))
SOTA
Two Wrongs Don't Make a Right: Combating Confirmation Bias in Learning with Label Noise arXiv:2112.02960
Image Classificationonmini WebVision 1.0
ImageNet Top-1 Accuracy· 2021-12-06
75.48
best: 82.56 (LRA-diffusion (CLIP ViT))
Two Wrongs Don't Make a Right: Combating Confirmation Bias in Learning with Label Noise arXiv:2112.02960
Image Classificationonmini WebVision 1.0
ImageNet Top-5 Accuracy· 2021-12-06
93.76
best: 97.24 (RTE (Inception-ResNet-v2))
Two Wrongs Don't Make a Right: Combating Confirmation Bias in Learning with Label Noise arXiv:2112.02960