Image Classification on VLCS (test)
81.59Average AccuracyWATT-S
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| WATT-SVisual Encoder=ViT-B/162024.06 | 81.59 | 99.36 | 75.16 | 83.24 | 68.59 | — | — | |
| WATT-PVisual Encoder=ViT-B/162024.06 | 81.05 | 99.36 | 74.75 | 82.53 | 67.55 | — | — | |
| CLIPVisual Encoder=ViT-B/162024.06 | 80.96 | 99.43 | 71.74 | 84.9 | 67.75 | — | — | |
| CLIPARTTVisual Encoder=ViT-B/162024.06 | 80.89 | 99.43 | 71.67 | 84.73 | 67.74 | — | — | |
| TENTVisual Encoder=ViT-B/162024.06 | 80.85 | 99.43 | 71.57 | 85.1 | 67.31 | — | — | |
| Core-tuningPre-training=MoCo-v2, Fine-tuning=Core-tuning, Backbone=ResNet-502021.02 | 80.34 | 98.5 | 81.53 | 73.15 | 68.19 | — | — | |
| TENTVisual Encoder=ViT-L/142024.06 | 80.2 | 99.43 | 68.23 | 84.08 | 69.07 | — | — | |
| CLIPVisual Encoder=ViT-L/142024.06 | 80.18 | 99.43 | 68.06 | 83.99 | 69.22 | — | — | |
| CLIPARTTVisual Encoder=ViT-L/142024.06 | 80.13 | 99.43 | 67.89 | 83.89 | 69.32 | — | — | |
| WATT-SVisual Encoder=ViT-L/142024.06 | 79.38 | 99.51 | 72.21 | 83.02 | 62.76 | — | — | |
| MoCo-v2 (CE-Con)Pre-training=MoCo-v2, Fine-tuning=CE-Con, Backbone=ResNet-502021.02 | 77.67 | 95.94 | 73.57 | 69.31 | 67.76 | — | — | |
| Supervised (CE)Pre-training=Supervised, Fine-tuning=Cross-Entropy (CE), Backbone=ResNet-502021.02 | 76.56 | 98.41 | 75.45 | 68.55 | 63.81 | — | — | |
| StableFDG2023.11 | 75.6 | 98.5 | 69.43 | 74.4 | 60.07 | — | — | |
| RSCBackbone=AlexNet2022.12 | 75.43 | 97.61 | 68.32 | 73.93 | 61.86 | — | — | |
| RotationBackbone=Alexnet, Evaluation Protocol=max target accuracy over training period2020.07 | 75.41 | 98.4 | 67.4 | 73.03 | 62.8 | — | — | |
| LRDGBackbone=AlexNet2022.12 | 75.15 | 96.85 | 69.27 | 68.95 | 65.53 | — | — | |
| JigsawBackbone=Alexnet, Evaluation Protocol=max target accuracy over training period2020.07 | 74.96 | 98.27 | 66.53 | 73.61 | 61.44 | — | — | |
| Jigsaw+RotationBackbone=Alexnet, Evaluation Protocol=max target accuracy over training period2020.07 | 74.94 | 98.1 | 68.87 | 72.6 | 60.2 | — | — | |
| DSU2023.11 | 74.76 | 96.87 | 68.97 | 72.97 | 60.23 | — | — | |
| MetaVIBBackbone=AlexNet2022.12 | 74.54 | 97.37 | 67.85 | 70.28 | 62.66 | — | — | |
| TSDBackbone=ResNet-502023.03 | 74.52 | — | — | — | — | — | — | |
| ERBackbone=AlexNet2022.12 | 74.38 | 96.92 | 69.1 | 73.24 | 58.26 | — | — | |
| MixStyle2023.11 | 74.15 | 95.2 | 66.73 | 73.9 | 60.77 | — | — | |
| FedDG2023.11 | 74.15 | 96.2 | 67.3 | 72.4 | 60.7 | — | — | |
| MASFBackbone=Alexnet, Evaluation Protocol=max target accuracy over training period2020.07 | 74.11 | 94.78 | 67.64 | 69.14 | 64.9 | — | — | |
| MASFBackbone=AlexNet2022.12 | 74.11 | 94.78 | 67.64 | 69.14 | 64.9 | — | — | |
| ERMBackbone=ResNet-502023.03 | 74.01 | — | — | — | — | — | — | |
| TPTVisual Encoder=ViT-L/142024.06 | 74.01 | 97.86 | 69.49 | 76.16 | 52.54 | — | — | |
| T3ABackbone=ResNet-502023.03 | 73.98 | — | — | — | — | — | — | |
| FedBN2023.11 | 73.97 | 94.34 | 69.04 | 69.89 | 62.61 | — | — | |
| MoCo-v2 (CE)Pre-training=MoCo-v2, Fine-tuning=Cross-Entropy (CE), Backbone=ResNet-502021.02 | 73.94 | 94.96 | 64.98 | 68.96 | 66.87 | — | — | |
| LAMEBackbone=ResNet-502023.03 | 73.94 | — | — | — | — | — | — | |
| MMLDBackbone=Alexnet2020.07 | 73.88 | 96.66 | 68.13 | 71.96 | 58.77 | — | — | |
| RotationBackbone=Alexnet, alpha=0.9, beta=0.62020.07 | 73.88 | 97.3 | 65.97 | 71.93 | 60.3 | — | — | |
| FedSR2023.11 | 73.46 | 92.6 | 68.3 | 72.15 | 60.8 | — | — | |
| DeepAllBackbone=Alexnet, Reference=MMLD [53]2020.07 | 73.39 | 95.89 | 67.76 | 72.01 | 57.88 | — | — | |
| CCST2023.11 | 73.38 | 96.7 | 65 | 71.4 | 60.4 | — | — | |
| JigsawBackbone=Alexnet, alpha=0.5, beta=0.82020.07 | 73.33 | 96.46 | 64.4 | 72.95 | 59.51 | — | — | |
| RExBackbone=AlexNet (ImageNet pre-trained), Selection Criterion=Highest validation accuracy2020.03 | 73.3 | 96.72 | 63.68 | 72.41 | 60.4 | — | — | |
| JiGenBackbone=AlexNet2022.12 | 73.19 | 96.93 | 64.3 | 70.62 | 60.9 | — | — | |
| FedAvg2023.11 | 73.18 | 93.65 | 65.4 | 72.55 | 61.1 | — | — | |
| Jigsaw+RotationBackbone=Alexnet, alpha_J=0.9, alpha_R=0.5, beta=0.72020.07 | 73.15 | 96.3 | 66.37 | 70.73 | 59.2 | — | — | |
| Epi-FCRBackbone=Alexnet2020.07 | 72.9 | 94.1 | 65.9 | 67.1 | 64.3 | — | — | |
| Epi-FCRBackbone=AlexNet2022.12 | 72.9 | 94.1 | 65.9 | 67.1 | 64.3 | — | — | |
| JigsawBackbone=AlexNet (ImageNet pre-trained), Selection Criterion=Highest validation accuracy2020.03 | 72.71 | 96.46 | 63.84 | 70.49 | 60.06 | — | — | |
| IRMBackbone=AlexNet (ImageNet pre-trained), Selection Criterion=Highest validation accuracy2020.03 | 72.54 | 95.99 | 62.85 | 71.71 | 59.61 | — | — | |
| TPTVisual Encoder=ViT-B/162024.06 | 72.53 | 97.62 | 71.56 | 71.17 | 49.77 | — | — | |
| DeepAllBackbone=Alexnet, Reference=This Paper (Baseline)2020.07 | 72.49 | 96.15 | 63.92 | 70.84 | 59.05 | — | — | |
| StylizedBackbone=AlexNet2022.12 | 72.31 | 96.86 | 63.42 | 68.18 | 60.77 | — | — | |
| DeepAllBackbone=Alexnet, Reference=MASF [25]2020.07 | 72.19 | 92.86 | 64.11 | 68.67 | 63.1 | — | — | |
| TFBackbone=Alexnet2020.07 | 72.11 | 93.63 | 61.32 | 69.99 | 63.49 | — | — | |
| DeepAllBackbone=Alexnet, Reference=TF [44]2020.07 | 72.02 | 93.4 | 64.16 | 68.41 | 62.11 | — | — | |
| BaselineBackbone=AlexNet2022.12 | 72.01 | 96.17 | 63.78 | 66.27 | 61.81 | — | — | |
| ERMBackbone=AlexNet (ImageNet pre-trained), Selection Criterion=Highest validation accuracy2020.03 | 71.56 | 94.76 | 61.92 | 69.03 | 60.55 | — | — | |
| DeepAllBackbone=Alexnet, Reference=Epi-FCR [45]2020.07 | 71.2 | 93.1 | 65.8 | 65.4 | 60.6 | — | — | |
| DeepAllBackbone=Alexnet, Reference=D-SAM [22]2020.07 | 71.08 | 94.95 | 65.87 | 66.06 | 57.45 | — | — | |
| SLRCBackbone=Alexnet2020.07 | 70.97 | 92.76 | 63.54 | 65.25 | 62.34 | — | — | |
| CIDDGBackbone=AlexNet2022.12 | 69.59 | 88.83 | 62.1 | 64.38 | 63.06 | — | — | |
| TentBackbone=ResNet-502023.03 | 69.2 | — | — | — | — | — | — | |
| PLBackbone=ResNet-502023.03 | 68.52 | — | — | — | — | — | — | |
| D-SAMBackbone=Alexnet2020.07 | 67.03 | 91.75 | 60.84 | 58.59 | 56.95 | — | — | |
| DeepAllBackbone=Alexnet, Reference=SLRC [20]2020.07 | 65.46 | 86.67 | 57.86 | 59.1 | 58.2 | — | — | |
| SHOT-IMBackbone=ResNet-502023.03 | 65.23 | — | — | — | — | — | — | |
| ETABackbone=ResNet-502023.03 | 64.79 | — | — | — | — | — | — | |
| BNBackbone=ResNet-502023.03 | 64.78 | — | — | — | — | — | — | |
| DenseModel=ResNet-182024.04 | — | — | — | — | — | 80.89 | — | |
| DenseModel=MobileNet-V22024.04 | — | — | — | — | — | 81.83 | — | |
| DenseModel=Swin-T2024.04 | — | — | — | — | — | 86.58 | — | |
| EGPModel=ResNet-182024.04 | — | — | — | — | — | 74.28 | — | |
| EGPModel=MobileNet-V22024.04 | — | — | — | — | — | 45.85 | — | |
| EGPModel=Swin-T2024.04 | — | — | — | — | — | 82.95 | — | |
| Group lassoModel=ResNet-182024.04 | — | — | — | — | — | 67.85 | — | |
| Group lassoModel=MobileNet-V22024.04 | — | — | — | — | — | 78.84 | — | |
| Group lassoModel=Swin-T2024.04 | — | — | — | — | — | 84.81 | — | |
| IMPModel=ResNet-182024.04 | — | — | — | — | — | 74.09 | — | |
| IMPModel=MobileNet-V22024.04 | — | — | — | — | — | 79.43 | — | |
| IMPModel=Swin-T2024.04 | — | — | — | — | — | 80.06 | — | |
| NEPENTHEModel=ResNet-182024.04 | — | — | — | — | — | 78.38 | — | |
| NEPENTHEModel=MobileNet-V22024.04 | — | — | — | — | — | 80.06 | — | |
| NEPENTHEModel=Swin-T2024.04 | — | — | — | — | — | 85.27 | — | |
| Smallest gradientsModel=ResNet-182024.04 | — | — | — | — | — | 46.13 | — | |
| Smallest gradientsModel=MobileNet-V22024.04 | — | — | — | — | — | 46.13 | — | |
| Smallest gradientsModel=Swin-T2024.04 | — | — | — | — | — | 84.15 | — | |
| Smallest weightsModel=ResNet-182024.04 | — | — | — | — | — | 46.13 | — | |
| Smallest weightsModel=MobileNet-V22024.04 | — | — | — | — | — | 6.43 | — | |
| Smallest weightsModel=Swin-T2024.04 | — | — | — | — | — | 84.62 | — |