Image Classification on WebVision mini (val)
81.84Top-1 AccuracyRobust LR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Robust LR2021.12 | 81.84 | 94.12 | |
| PGDF2021.12 | 81.47 | 94.03 | |
| C2DArchitecture=ResNet-50, base_method=DivideMix with SimCLR2021.03 | 79.42 | 92.32 | |
| TSCSI_IDN2022.07 | 79.36 | 93.64 | |
| NGC2021.12 | 79.16 | 91.84 | |
| NGC2022.07 | 79.16 | 91.84 | |
| NGCBackbone=ResNet-182023.03 | 79.1 | 91.8 | |
| TCLBackbone=ResNet-182023.03 | 79.1 | 92.3 | |
| LongReMix2021.12 | 78.92 | 92.32 | |
| MOIT+Selection Criterion=Best2020.12 | 78.76 | — | |
| MOIT+Selection Criterion=Last2020.12 | 78.72 | — | |
| MOITSelection Criterion=Best2020.12 | 78.36 | — | |
| MOITBackbone=ResNet-182023.03 | 77.9 | 91.9 | |
| ELR+Architecture=InceptionResNetV2, Training Dataset=mini WebVision2020.06 | 77.78 | 91.68 | |
| ELR+2021.12 | 77.78 | 91.68 | |
| ELR+Architecture=Inception-ResNet-v22021.03 | 77.78 | 91.68 | |
| ELR+Backbone=InceptionResNetV2, Training Dataset=mini WebVision, Mixup Data Augmentation=true, Two Networks=true, Weight Averaging=true2021.08 | 77.78 | 91.68 | |
| MOITSelection Criterion=Last2020.12 | 77.76 | — | |
| CARBackbone=InceptionResNetV2, Training Dataset=mini WebVision2021.08 | 77.41 | 92.25 | |
| DivideMixTraining dataset=(mini) WebVision2020.02 | 77.32 | 91.64 | |
| DivideMixArchitecture=InceptionResNetV2, Training Dataset=mini WebVision2020.06 | 77.32 | 91.64 | |
| DivideMixTraining dataset=(mini) WebVision2021.02 | 77.32 | 91.64 | |
| DivideMix2021.12 | 77.32 | 91.64 | |
| DivideMixArchitecture=Inception-ResNet-v22021.03 | 77.32 | 91.64 | |
| DivideMix2022.07 | 77.32 | 91.64 | |
| DivideMix2021.12 | 77.32 | 91.64 | |
| DivideMixBackbone=InceptionResNetV2, Training Dataset=mini WebVision, Mixup Data Augmentation=true, Two Networks=true, Weight Averaging=true2021.08 | 77.32 | 91.64 | |
| DivideMixBackbone=ResNet-182023.03 | 77.3 | 91.6 | |
| F-DivideMixTraining dataset=(mini) WebVision, Note=Reduced batch size, no hyperparameter tuning2021.02 | 77.28 | 91.44 | |
| DivideMixArchitecture=ResNet-50, implementation_source=acquired by authors based on published code2021.03 | 76.32 | 90.65 | |
| RRLBackbone=ResNet-182023.03 | 76.3 | 91.5 | |
| ELRArchitecture=InceptionResNetV2, Training Dataset=mini WebVision2020.06 | 76.26 | 91.26 | |
| ELRBackbone=InceptionResNetV2, Training Dataset=mini WebVision2021.08 | 76.26 | 91.26 | |
| ELRBackbone=ResNet-182023.03 | 76.2 | 91.2 | |
| DMixSelection Criterion=Best2020.12 | 76.08 | — | |
| MentorMixArchitecture=Inception-ResNet-v22021.03 | 76 | 90.2 | |
| FINETraining dataset=(mini) WebVision, Note=Reduced batch size, no hyperparameter tuning2021.02 | 75.24 | 90.28 | |
| NCT2021.12 | 75.16 | 90.77 | |
| MixSelection Criterion=Best2020.12 | 74.96 | — | |
| DMixSelection Criterion=Last2020.12 | 74.64 | — | |
| ODDArchitecture=Inception-ResNet-v22021.03 | 74.6 | 90.6 | |
| MixSelection Criterion=Last2020.12 | 73.76 | — | |
| ELRSelection Criterion=Best2020.12 | 73 | — | |
| CRUSTTraining dataset=(mini) WebVision2021.02 | 72.4 | 89.56 | |
| ELRSelection Criterion=Last2020.12 | 71.88 | — | |
| Iterative-CVTraining dataset=(mini) WebVision2020.02 | 65.24 | 85.34 | |
| Iterative-CVArchitecture=InceptionResNetV2, Training Dataset=mini WebVision2020.06 | 65.24 | 85.34 | |
| Iterative-CV2022.07 | 65.24 | 85.34 | |
| Iterative-CV2021.12 | 65.24 | 85.34 | |
| Iterative-CVBackbone=InceptionResNetV2, Training Dataset=mini WebVision2021.08 | 65.24 | 85.34 | |
| Iterative-CVArchitecture=Inception-ResNet-v22021.03 | 65.2 | 85.3 | |
| Iterative-CVBackbone=ResNet-182023.03 | 65.2 | 85.3 | |
| Co-teachingTraining dataset=(mini) WebVision2020.02 | 63.58 | 85.2 | |
| Co-teachingArchitecture=InceptionResNetV2, Training Dataset=mini WebVision2020.06 | 63.58 | 85.2 | |
| Co-teaching2022.07 | 63.58 | 85.2 | |
| Co-teaching2021.12 | 63.58 | 85.2 | |
| Co-teachingBackbone=InceptionResNetV2, Training Dataset=mini WebVision2021.08 | 63.58 | 85.2 | |
| Co-teachingBackbone=ResNet-182023.03 | 63.5 | 85.2 | |
| MentorNetTraining dataset=(mini) WebVision2020.02 | 63 | 81.4 | |
| MentorNetArchitecture=InceptionResNetV2, Training Dataset=mini WebVision2020.06 | 63 | 81.4 | |
| MentorNet2022.07 | 63 | 81.4 | |
| MentorNet2021.12 | 63 | 81.4 | |
| MentorNetBackbone=ResNet-182023.03 | 63 | 81.4 | |
| MentorNetBackbone=InceptionResNetV2, Training Dataset=mini WebVision2021.08 | 63 | 81.4 | |
| D2LTraining dataset=(mini) WebVision2020.02 | 62.68 | 84 | |
| D2LArchitecture=InceptionResNetV2, Training Dataset=mini WebVision2020.06 | 62.68 | 84 | |
| D2L2022.07 | 62.68 | 84 | |
| D2L2021.12 | 62.68 | 84 | |
| D2LBackbone=InceptionResNetV2, Training Dataset=mini WebVision2021.08 | 62.68 | 84 | |
| D2LBackbone=ResNet-182023.03 | 62.6 | 84 | |
| DecouplingTraining dataset=(mini) WebVision2020.02 | 62.54 | 84.74 | |
| Decoupling2022.07 | 62.54 | 84.74 | |
| Decoupling2021.12 | 62.54 | 84.74 | |
| DecouplingBackbone=ResNet-182023.03 | 62.5 | 84.7 | |
| F-correctionTraining dataset=(mini) WebVision2020.02 | 61.12 | 82.68 | |
| F-correction2022.07 | 61.12 | 82.68 | |
| F-correction2021.12 | 61.12 | 82.68 | |
| ForwardBackbone=ResNet-182023.03 | 61.1 | 82.6 |