Image Classification on ImageNet-1K Noisy 1.0 (val)
69.4AccuracyInfoBatch + AlignPrune
Evaluation results
| Method | Links | ||
|---|---|---|---|
| InfoBatch + AlignPruneModel Architecture=DeiT, Model Scale=Base2026.04 | 69.4 | 0.5 | |
| Full-trainingModel Architecture=DeiT, Model Scale=Base2026.04 | 68.9 | — | |
| InfoBatch + AlignPruneModel Architecture=Swin, Model Scale=Base2026.04 | 68.2 | 0.4 | |
| InfoBatchModel Architecture=DeiT, Model Scale=Base2026.04 | 68.1 | -0.8 | |
| InfoBatch + AlignPruneModel Architecture=ConvNeXt, Model Scale=Base2026.04 | 67.8 | 0.5 | |
| Full-trainingModel Architecture=Swin, Model Scale=Base2026.04 | 67.8 | — | |
| Full-trainingModel Architecture=ConvNeXt, Model Scale=Base2026.04 | 67.3 | — | |
| InfoBatch + AlignPruneModel Architecture=ConvNeXt, Model Scale=Tiny2026.04 | 67.2 | 0.8 | |
| InfoBatchModel Architecture=Swin, Model Scale=Base2026.04 | 66.9 | -0.9 | |
| InfoBatchModel Architecture=ConvNeXt, Model Scale=Base2026.04 | 66.8 | -0.5 | |
| Full-trainingModel Architecture=ConvNeXt, Model Scale=Tiny2026.04 | 66.4 | — | |
| InfoBatchModel Architecture=ConvNeXt, Model Scale=Tiny2026.04 | 65.8 | -0.6 | |
| InfoBatch + AlignPruneModel Architecture=Swin, Model Scale=Tiny2026.04 | 65.6 | 0.6 | |
| InfoBatch + AlignPruneModel Architecture=DeiT, Model Scale=Small2026.04 | 65.5 | 0.8 | |
| Full-trainingModel Architecture=Swin, Model Scale=Tiny2026.04 | 65 | — | |
| Full-trainingModel Architecture=DeiT, Model Scale=Small2026.04 | 64.7 | — | |
| InfoBatchModel Architecture=Swin, Model Scale=Tiny2026.04 | 64.3 | -0.7 | |
| InfoBatchModel Architecture=DeiT, Model Scale=Small2026.04 | 64.1 | -0.6 | |
| InfoBatchModel Architecture=Mean2026.04 | — | -0.7 | |
| InfoBatch + AlignPruneModel Architecture=Mean2026.04 | — | 0.6 |