Clip
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Image Classification | CLIP benchmark (test) | Accuracy93 | 72 | |
| Deferral-advice | CLIP prompt-escalation (val) | True Deferral-Advice Loss0.337 | 66 | |
| Membership Inference | CLIP image-text (train) | Precision91.68 | 36 | |
| Image Classification | CLIP Zero-shot Evaluation Suite (10 datasets) | Cars Accuracy89.6 | 16 | |
| Multilabel Classification | CLIP (test) | Micro F180.9 | 12 | |
| Image Classification | CLIP Classification Suite | CIFAR-10 Accuracy97.2 | 11 | |
| Classification | CLIP 21-dataset Zero-Shot standard (test val) | ImageNet Accuracy45 | 11 | |
| Classification | CLIP 200 random samples (test) | Macro F1 Score0.24 | 6 | |
| Lyric Intelligibility Prediction | CLIP (evaluation) | RMSE (%)27.07 | 3 | |
| Lyric Intelligibility Prediction | CLIP (val) | RMSE (%)27.13 | 3 | |
| Image Denoising | Clip300 sigma=60 | Average PSNR (dB)25.51 | 3 | |
| Image Denoising | Clip300 sigma=15 | Average PSNR (dB)31.68 | 3 | |
| Multi-shot Backdoor Classification | CLIP Multi-shot Downstream | BA70.27 | 2 |
Showing 13 of 13 rows