ResearchBenchmarksImage Classification on ImageNet-100 (80-20 split)For AI assistantsMarkdownContent indexImage Classification on ImageNet-100 (80-20 split)Follow69Accuracy (R)MAS*20.84833.34945.8558.351Mar 18, 2024Evaluation resultsMethodMethodLinksAccuracy (R)Harmonic MeanF1 ScoreOverall AccuracyMAS*Backbone=ViTBackbone=ViT2024.036957.112.213.7L2*Backbone=ViTBackbone=ViT2024.036847.624.330GS-LORABackbone=ViTBackbone=ViT2024.0366.961.73.60.8EWC*Backbone=ViTBackbone=ViT2024.0364.955.612.213.9Pre-trainBackbone=ViTBackbone=ViT2024.0364.1—60.9—RetrainBackbone=ViTBackbone=ViT2024.0322.7330.70