ResearchBenchmarksImage Classification on ImageNet (val) (Accuracy, Throughput, and MACs)For AI assistantsMarkdownContent indexImage Classification on ImageNet (val) (Accuracy, Throughput, and MACs)Follow83.3Top-1 AccuracyLV-ViT-S82.46882.68482.983.116Feb 16, 2022Evaluation resultsMethodMethodLinksTop-1 AccuracyTop-5 AccuracyThroughput (images/s)MACs (G)LV-ViT-SKeep rate=1.0Keep rate=1.02022.0283.3—2,1126.6EViTKeep rate=0.7, Backbon...Keep rate=0.7, Backbone=LV-ViT-S2022.028396.32,9544.7EViTKeep rate=0.5, Backbon...Keep rate=0.5, Backbone=LV-ViT-S2022.0282.596.23,6033.9