GPaCo (ViT-L)

Reported on 6 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.

Methodology3 results

Domain AdaptationonImageNet-R
Top-1 Error Rate· 2022-09-26
39.7
best: 3.9 (Model soups (BASIC-L))
Generalized Parametric Contrastive Learning arXiv:2209.12400
Domain AdaptationonImageNet-C
mean Corruption Error (mCE)· 2022-09-26
39
best: 22 (EfficientNet-L2+RPL)
Generalized Parametric Contrastive Learning arXiv:2209.12400
Domain AdaptationonImageNet-Sketch
Top-1 accuracy· 2022-09-26
48.3
best: 77.18 (Model soups (BASIC-L))
Generalized Parametric Contrastive Learning arXiv:2209.12400

Computer Vision3 results

Domain GeneralizationonImageNet-R
Top-1 Error Rate· 2022-09-26
39.7
best: 3.9 (Model soups (BASIC-L))
Generalized Parametric Contrastive Learning arXiv:2209.12400
Domain GeneralizationonImageNet-C
mean Corruption Error (mCE)· 2022-09-26
39
best: 28.2 (DINOv2 (ViT-g/14, frozen model, linear eval))
Generalized Parametric Contrastive Learning arXiv:2209.12400
Domain GeneralizationonImageNet-Sketch
Top-1 accuracy· 2022-09-26
48.3
best: 77.18 (Model soups (BASIC-L))
Generalized Parametric Contrastive Learning arXiv:2209.12400