Domain Generalization Using Large Pretrained Models with Mixture-of-Adapters

Open paper
Benchmark Model Rank Results
domain-generalization-on-domainnetMoA (OpenCLIP, ViT-B/16)#5Average Accuracy: 62.7
domain-generalization-on-office-homeMoA (OpenCLIP, ViT-B/16)#1Average Accuracy: 90.6
domain-generalization-on-pacs-2MoA (OpenCLIP, ViT-B/16)#5Average Accuracy: 97.4
domain-generalization-on-terraincognitaMoA (OpenCLIP, ViT-B/16)#13Average Accuracy: 52.8
domain-generalization-on-vlcsMoA (OpenCLIP, ViT-B/16)#4Average Accuracy: 83.1

Rank counts only results with a code link.