Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/Stylized ImageNet (ResNet-50)

Stylized ImageNet (ResNet-50)

Reported on 6 benchmarks across 2 tasks · 1 paper · 4 SOTA

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· uses extra data· 2018-11-29
58.5
best: 3.9 (Model soups (BASIC-L))
SOTA
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness arXiv:1811.12231
Domain AdaptationonImageNet-C
mean Corruption Error (mCE)· uses extra data· 2018-11-29
69.3
best: 22 (EfficientNet-L2+RPL)
SOTA
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness arXiv:1811.12231
Domain AdaptationonImageNet-A
Top-1 accuracy %· uses extra data· 2018-11-29
2.3
best: 94.17 (Model soups (BASIC-L))
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness arXiv:1811.12231

Computer Vision3 results

Domain GeneralizationonImageNet-R
Top-1 Error Rate· uses extra data· 2018-11-29
58.5
best: 3.9 (Model soups (BASIC-L))
SOTA
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness arXiv:1811.12231
Domain GeneralizationonImageNet-C
mean Corruption Error (mCE)· uses extra data· 2018-11-29
69.3
best: 28.2 (DINOv2 (ViT-g/14, frozen model, linear eval))
SOTA
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness arXiv:1811.12231
Domain GeneralizationonImageNet-A
Top-1 accuracy %· uses extra data· 2018-11-29
2.3
best: 94.17 (Model soups (BASIC-L))
ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness arXiv:1811.12231