Weight Averaging Improves Knowledge Distillation under Domain Shift

Open paper
Benchmark Model Rank Results
domain-generalization-on-office-homeWAKD (DeiT-Ti)#27Average Accuracy: 70.5
domain-generalization-on-office-homeWAKD (Resnet-18)#35Average Accuracy: 66.7
domain-generalization-on-pacs-2WAKD (DeiT-Ti)#33Average Accuracy: 87.6
domain-generalization-on-pacs-2WAKD (Resnet-18)#41Average Accuracy: 86.6

Rank counts only results with a code link.