DrViT

Reported on 4 benchmarks across 2 tasks · 1 paper

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

Methodology2 results

Domain AdaptationonImageNet-C
mean Corruption Error (mCE)· 2021-11-20
46.22
best: 22 (EfficientNet-L2+RPL)
Discrete Representations Strengthen Vision Transformer Robustness arXiv:2111.10493
Domain AdaptationonImageNet-Sketch
Top-1 accuracy· uses extra data· 2021-11-20
44.72
best: 77.18 (Model soups (BASIC-L))
Discrete Representations Strengthen Vision Transformer Robustness arXiv:2111.10493

Computer Vision2 results

Domain GeneralizationonImageNet-C
mean Corruption Error (mCE)· 2021-11-20
46.22
best: 28.2 (DINOv2 (ViT-g/14, frozen model, linear eval))
Discrete Representations Strengthen Vision Transformer Robustness arXiv:2111.10493
Domain GeneralizationonImageNet-Sketch
Top-1 accuracy· uses extra data· 2021-11-20
44.72
best: 77.18 (Model soups (BASIC-L))
Discrete Representations Strengthen Vision Transformer Robustness arXiv:2111.10493