Multi-task Image Classification on 8-task vision benchmark
95.8Average AccuracyFine-tuned
Evaluation results
| Method | Links | |
|---|---|---|
| Fine-tunedBackbone=ViT-L/142026.03 | 95.8 | |
| Fine-tunedBackbone=ViT-B/162026.03 | 94.6 | |
| ACE-MergingBackbone=ViT-L/142026.03 | 94.6 | |
| TSV-MBackbone=ViT-L/142026.03 | 93 | |
| Fine-tunedBackbone=ViT-B/322026.03 | 92.8 | |
| CARTBackbone=ViT-L/142026.03 | 92.6 | |
| ACE-MergingBackbone=ViT-B/162026.03 | 90.6 | |
| TSV-MBackbone=ViT-B/162026.03 | 89 | |
| CARTBackbone=ViT-B/162026.03 | 88.3 | |
| ACE-MergingBackbone=ViT-B/322026.03 | 87.9 | |
| Consensus TABackbone=ViT-L/142026.03 | 86.3 | |
| TSV-MBackbone=ViT-B/322026.03 | 85.9 | |
| Task ArithmeticBackbone=ViT-L/142026.03 | 84.9 | |
| CARTBackbone=ViT-B/322026.03 | 84.7 | |
| Weight AveragingBackbone=ViT-L/142026.03 | 79.6 | |
| Consensus TABackbone=ViT-B/162026.03 | 79.4 | |
| Task ArithmeticBackbone=ViT-B/162026.03 | 75.4 | |
| Consensus TABackbone=ViT-B/322026.03 | 75 | |
| Weight AveragingBackbone=ViT-B/162026.03 | 72.2 | |
| Task ArithmeticBackbone=ViT-B/322026.03 | 70.8 | |
| Weight AveragingBackbone=ViT-B/322026.03 | 66.3 | |
| Zero-shotBackbone=ViT-L/142026.03 | 64.7 | |
| Zero-shotBackbone=ViT-B/162026.03 | 55.3 | |
| Zero-shotBackbone=ViT-B/322026.03 | 48.3 |