ResearchBenchmarksFine-Grained Visual Categorization on NABirdsFine-Grained Visual Categorization on NABirdsFollow91.2AccuracySR-GNN89.84890.19990.5590.901Sep 5, 2022Evaluation resultsMethodMethodLinksAccuracyParameters (M)GFLOPs (B)SR-GNNInput Size=224x224Input Size=224x2242022.0991.230.99.8TrnFGTransformer=ViT-B-16,...Transformer=ViT-B-16, Input Size=448x4482022.0990.8——ViTTransformer=ViT-B-16,...Transformer=ViT-B-16, Input Size=448x4482022.0989.98655.4