Image Classification on CLIP 8-task suite
94.3Average Top-1 AccuracyFine-tuned
Evaluation results
| Method | Links | |
|---|---|---|
| Fine-tunedBackbone=CLIP-ViT-L/142026.05 | 94.3 | |
| TIES-Merging w/ E-PMQBackbone=CLIP-ViT-L/142026.05 | 86 | |
| Task Arithmetic w/ E-PMQBackbone=CLIP-ViT-L/142026.05 | 85.9 | |
| TIES-MergingBackbone=CLIP-ViT-L/142026.05 | 84 | |
| TIES-Merging w/ GPTQBackbone=CLIP-ViT-L/142026.05 | 81.8 | |
| TIES-Merging w/ RTNBackbone=CLIP-ViT-L/142026.05 | 81.2 | |
| Task ArithmeticBackbone=CLIP-ViT-L/142026.05 | 80.7 | |
| Task Arithmetic w/ GPTQBackbone=CLIP-ViT-L/142026.05 | 78.3 | |
| Task Arithmetic w/ RTNBackbone=CLIP-ViT-L/142026.05 | 77.3 | |
| TIES-Merging w/ AWQBackbone=CLIP-ViT-L/142026.05 | 73.4 | |
| Task Arithmetic w/ AWQBackbone=CLIP-ViT-L/142026.05 | 66.9 | |
| Pre-trainedBackbone=CLIP-ViT-L/142026.05 | 64.6 |