CoDiM-Self (Inception-ResNet-v2)

Reported on 4 benchmarks across 1 task · 1 paper · 1 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-5 Accuracy· 2021-11-23
93.52
best: 94.84 (PSSCL (130 epochs))
SOTA
CoDiM: Learning with Noisy Labels via Contrastive Semi-Supervised Learning arXiv:2111.11652
Image Classificationonmini WebVision 1.0
ImageNet Top-1 Accuracy· 2021-11-23
77.24
best: 82.56 (LRA-diffusion (CLIP ViT))
CoDiM: Learning with Noisy Labels via Contrastive Semi-Supervised Learning arXiv:2111.11652
Image Classificationonmini WebVision 1.0
ImageNet Top-5 Accuracy· 2021-11-23
92.48
best: 97.24 (RTE (Inception-ResNet-v2))
CoDiM: Learning with Noisy Labels via Contrastive Semi-Supervised Learning arXiv:2111.11652
Image Classificationonmini WebVision 1.0
Top-1 Accuracy· 2021-11-23
80.12
best: 84.16 (LRA-diffusion (CLIP ViT))
CoDiM: Learning with Noisy Labels via Contrastive Semi-Supervised Learning arXiv:2111.11652