Human Perceptual Alignment on NIGHT (test)
96.2Overall ScoreEnsemble
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| EnsembleModel Class=Tuned LoRA, Model Type=ViT B/16, Feature=Mixed2023.06 | 96.2 | 96.6 | 95.5 | |
| OpenCLIPModel Class=Tuned LoRA, Model Type=ViT B/32, Feature=Embedding2023.06 | 95.5 | 96.5 | 94.1 | |
| DINOModel Class=Tuned LoRA, Model Type=ViT B/16, Feature=CLS2023.06 | 94.6 | 94.6 | 94.5 | |
| CLIPModel Class=Tuned LoRA, Model Type=ViT B/32, Feature=Embedding2023.06 | 93.9 | 94 | 93.6 | |
| EnsembleModel Class=Tuned MLP, Model Type=ViT B/16, Feature=Mixed2023.06 | 93.4 | 94.2 | 92.2 | |
| DINOModel Class=Tuned MLP, Model Type=ViT B/16, Feature=CLS2023.06 | 91.2 | 91.8 | 90.3 | |
| EnsembleModel Class=Base Models, Model Type=ViT B/16, Feature=Mixed2023.06 | 90.8 | 91.6 | 89.8 | |
| MAEModel Class=Tuned LoRA, Model Type=ViT B/16, Feature=CLS2023.06 | 90.5 | 90.6 | 90.3 | |
| DINOModel Class=Base Models, Model Type=ViT B/16, Feature=CLS2023.06 | 90.1 | 90.6 | 89.5 | |
| OpenCLIPModel Class=Tuned MLP, Model Type=ViT B/32, Feature=Embedding2023.06 | 89.9 | 91 | 88.5 | |
| DINOModel Class=Base Models, Model Type=ViT S/16, Feature=CLS2023.06 | 89.6 | 90.2 | 88.8 | |
| DINOModel Class=Base Models, Model Type=ViT S/8, Feature=CLS2023.06 | 89 | 89.7 | 88 | |
| DINOModel Class=Base Models, Model Type=ViT B/8, Feature=CLS2023.06 | 88.6 | 88.6 | 88.5 | |
| OpenCLIPModel Class=Base Models, Model Type=ViT B/32, Feature=Embedding2023.06 | 87.6 | 87.5 | 87.6 | |
| MAEModel Class=Tuned MLP, Model Type=ViT B/16, Feature=CLS2023.06 | 87.6 | 87.3 | 88.1 | |
| CLIPModel Class=Tuned MLP, Model Type=ViT B/32, Feature=Embedding2023.06 | 87.3 | 88.2 | 86.2 | |
| OpenCLIPModel Class=Base Models, Model Type=ViT B/16, Feature=Embedding2023.06 | 87.1 | 87.8 | 86.2 | |
| DISTSModel Class=Prior-Learned Metrics, Model Type=VGG162023.06 | 86 | 87.1 | 84.5 | |
| OpenCLIPModel Class=Base Models, Model Type=ViT L/14, Feature=Embedding2023.06 | 85.9 | 86.7 | 84.9 | |
| MAEModel Class=Base Models, Model Type=ViT H/14, Feature=CLS2023.06 | 83.5 | 83.2 | 83.8 | |
| CLIPModel Class=Base Models, Model Type=ViT B/32, Feature=Embedding2023.06 | 83.1 | 83.8 | 82.1 | |
| MAEModel Class=Base Models, Model Type=ViT L/16, Feature=CLS2023.06 | 82.7 | 82.4 | 83 | |
| CLIPModel Class=Base Models, Model Type=ViT B/16, Feature=Embedding2023.06 | 82.2 | 82.6 | 81.7 | |
| MAEModel Class=Base Models, Model Type=ViT B/16, Feature=CLS2023.06 | 82.1 | 81 | 83.5 | |
| CLIPModel Class=Base Models, Model Type=ViT L/14, Feature=Embedding2023.06 | 81.8 | 83.3 | 79.8 | |
| LPIPSModel Class=Prior-Learned Metrics, Model Type=AlexNet-Linear2023.06 | 70.7 | 69.3 | 72.7 | |
| SSIMModel Class=Base Models2023.06 | 57.3 | 58.5 | 55.8 | |
| PSNRModel Class=Base Models2023.06 | 57.1 | 57.3 | 57 |