ResearchBenchmarksImage Classification on ImageNet-100 (60-20 split)For AI assistantsMarkdownContent indexImage Classification on ImageNet-100 (60-20 split)Follow74.1Accuracy (Rate)GS-LORA29.27640.91352.5564.187Mar 18, 2024Evaluation resultsMethodMethodLinksAccuracy (Rate)EntropyF1 ScoreOverall AccuracyGS-LORABackbone=ViTBackbone=ViT2024.0374.165.940.5EWC*Backbone=ViTBackbone=ViT2024.0372.753.820.77.8MAS*Backbone=ViTBackbone=ViT2024.0372.75420.47.7L2*Backbone=ViTBackbone=ViT2024.0370.14035.324.5Pre-trainBackbone=ViTBackbone=ViT2024.0369.2—63.3—RetrainBackbone=ViTBackbone=ViT2024.033141.12.20