Image Classification on ImageNet-1K Clean 1.0 (val)
77.8AccuracyFull-training
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Full-trainingModel Architecture=DeiT, Model Scale=Base2026.04 | 77.8 | — | |
| InfoBatchModel Architecture=DeiT, Model Scale=Base2026.04 | 77.8 | 0 | |
| InfoBatch + AlignPruneModel Architecture=DeiT, Model Scale=Base2026.04 | 77.8 | 0 | |
| InfoBatch + AlignPruneModel Architecture=Swin, Model Scale=Base2026.04 | 77.2 | 0.3 | |
| InfoBatchModel Architecture=Swin, Model Scale=Base2026.04 | 77.1 | 0.2 | |
| Full-trainingModel Architecture=Swin, Model Scale=Base2026.04 | 76.9 | — | |
| InfoBatch + AlignPruneModel Architecture=ConvNeXt, Model Scale=Base2026.04 | 75.7 | 0.1 | |
| Full-trainingModel Architecture=ConvNeXt, Model Scale=Base2026.04 | 75.6 | — | |
| InfoBatchModel Architecture=ConvNeXt, Model Scale=Base2026.04 | 75.4 | -0.2 | |
| InfoBatchModel Architecture=Swin, Model Scale=Tiny2026.04 | 73.6 | 0.1 | |
| InfoBatch + AlignPruneModel Architecture=Swin, Model Scale=Tiny2026.04 | 73.6 | 0.1 | |
| Full-trainingModel Architecture=Swin, Model Scale=Tiny2026.04 | 73.5 | — | |
| InfoBatch + AlignPruneModel Architecture=ConvNeXt, Model Scale=Tiny2026.04 | 73.4 | 0.1 | |
| Full-trainingModel Architecture=ConvNeXt, Model Scale=Tiny2026.04 | 73.3 | — | |
| InfoBatchModel Architecture=ConvNeXt, Model Scale=Tiny2026.04 | 73.2 | -0.1 | |
| InfoBatch + AlignPruneModel Architecture=DeiT, Model Scale=Small2026.04 | 72.1 | 0.2 | |
| Full-trainingModel Architecture=DeiT, Model Scale=Small2026.04 | 71.9 | — | |
| InfoBatchModel Architecture=DeiT, Model Scale=Small2026.04 | 71.5 | -0.4 | |
| InfoBatchModel Architecture=Mean2026.04 | — | -0.1 | |
| InfoBatch + AlignPruneModel Architecture=Mean2026.04 | — | 0.1 |