Domain Generalization On Imagenet Sketch

Metrics

Top-1 accuracy

Results

Performance results of various models on this benchmark

		Paper Title	Repository
Model soups (BASIC-L)	77.18	Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
Model soups (ViT-G/14)	74.24	Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
CAR-FT (CLIP, ViT-L/14@336px)	65.5	Context-Aware Robust Fine-Tuning	-
ConvNeXt-XL (Im21k, 384)	55.0	A ConvNet for the 2020s
CAFormer-B36 (IN21K, 384)	54.5	MetaFormer Baselines for Vision
LLE (ViT-H/14, MAE, Edge Aug)	53.39	A Whac-A-Mole Dilemma: Shortcuts Come in Multiples Where Mitigating One Amplifies Others
ConvFormer-B36 (IN21K, 384)	52.9	MetaFormer Baselines for Vision
CAFormer-B36 (IN21K)	52.8	MetaFormer Baselines for Vision
ConvFormer-B36 (IN21K)	52.7	MetaFormer Baselines for Vision
MAE (ViT-H, 448)	50.9	Masked Autoencoders Are Scalable Vision Learners
MAE+DAT (ViT-H)	50.03	Enhance the Visual Representation via Discrete Adversarial Training
GPaCo (ViT-L)	48.3	Generalized Parametric Contrastive Learning
Discrete Adversarial Distillation (ViT-B, 224)	46.1	Distilling Out-of-Distribution Robustness from Vision-Language Foundation Models
Pyramid Adversarial Training Improves ViT (Im21k)	46.03	Pyramid Adversarial Training Improves ViT Performance
SEER (RegNet10B)	45.6	Vision Models Are More Robust And Fair When Pretrained On Uncurated Images Without Supervision
DrViT	44.72	Discrete Representations Strengthen Vision Transformer Robustness
CAFormer-B36	42.5	MetaFormer Baselines for Vision
Pyramid Adversarial Training Improves ViT	41.04	Pyramid Adversarial Training Improves ViT Performance
ConvFormer-B36	39.5	MetaFormer Baselines for Vision
Sequencer2D-L	35.8	Sequencer: Deep LSTM for Image Classification

0 of 20 row(s) selected.

Command Palette

Domain Generalization On Imagenet Sketch

Metrics

Results