ResearchDatasetseight-task benchmarkFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsImage Classificationeight-task benchmark CLIP-ViT (minimum over eight per-task cells)76.3Worst-task Top-1 Accuracy15
Image Classificationeight-task benchmark CLIP-ViT (minimum over eight per-task cells)76.3Worst-task Top-1 Accuracy15