Image Classification on CIFAR-10 (Standard Accuracy)
100AccuracyRetrain
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Retrain2023.03 | 100 | — | — | — | — | — | — | |
| Finetune2023.03 | 100 | — | — | — | — | — | — | |
| Original Model2023.03 | 99.97 | — | — | — | — | — | — | |
| Mini-DeiT-B↑384#Params=44M2022.04 | 99.3 | — | — | — | — | — | — | |
| Boundary Shrink2023.03 | 99.24 | — | — | — | — | — | — | |
| DeiT-B⚗ ↑384#Params=88M2022.04 | 99.2 | — | — | — | — | — | — | |
| CMT-S# Params=25.1M, # FLOPs=4.04B, Pre-training=ImageNet, Fine-tuning=true2021.07 | 99.2 | — | — | — | — | — | — | |
| Heinsen RoutingBackbone=BEiT-large, Transformer Status=Frozen, Routing Iterations (n_iters)=22022.11 | 99.2 | — | — | — | — | — | — | |
| DeiT-B#Params=86M2022.04 | 99.1 | — | — | — | — | — | — | |
| DeiT-B↑384#Params=87M2022.04 | 99.1 | — | — | — | — | — | — | |
| DeiT-B⚗#Params=87M2022.04 | 99.1 | — | — | — | — | — | — | |
| DeiT-B# Params=85.8M, # FLOPs=17.6B, Pre-training=ImageNet, Fine-tuning=true2021.07 | 99.1 | — | — | — | — | — | — | |
| CeiT-S# Params=24.2M, # FLOPs=12.9B, Resolution=384, Pre-training=ImageNet, Fine-tuning=true2021.07 | 99.1 | — | — | — | — | — | — | |
| EfficientNet-B7#Params=66M2022.04 | 98.9 | — | — | — | — | — | — | |
| EfficientNet-B7# Params=64.0M, # FLOPs=37.2B, Resolution=600, Pre-training=ImageNet, Fine-tuning=true2021.07 | 98.9 | — | — | — | — | — | — | |
| EfficientNet-B5#Params=30M2022.04 | 98.7 | — | — | — | — | — | — | |
| ViT-Slimps#Params (M)=15.6, FLOPs (B)=3.32022.01 | 98.7 | — | — | — | — | — | — | |
| TNT-S# Params=23.8M, # FLOPs=17.3B, Resolution=384, Pre-training=ImageNet, Fine-tuning=true2021.07 | 98.7 | — | — | — | — | — | — | |
| DEIT-S#Params (M)=22.0, FLOPs (B)=4.62022.01 | 98.56 | — | — | — | — | — | — | |
| Random Labels2023.03 | 98.49 | — | — | — | — | — | — | |
| NVP-T#Params=6.9M2022.04 | 98.3 | — | — | — | — | — | — | |
| ISyNet-N32021.09 | 98.3 | — | — | — | — | — | — | |
| NViT-T#Params=6.4M2022.04 | 98.2 | — | — | — | — | — | — | |
| ResNet-34+2021.09 | 98.17 | — | — | — | — | — | — | |
| ResNet-50+2021.09 | 98.15 | — | — | — | — | — | — | |
| ViT-B/16#Params=86M2022.04 | 98.1 | — | — | — | — | — | — | |
| ViT-B/16# Params=85.8M, # FLOPs=17.6B, Resolution=384, Pre-training=ImageNet, Fine-tuning=true2021.07 | 98.1 | — | — | — | — | — | — | |
| ISyNet-N22021.09 | 98.07 | — | — | — | — | — | — | |
| ISyNet-N1-S22021.09 | 98.05 | — | — | — | — | — | — | |
| Boundary Expanding2023.03 | 98.03 | — | — | — | — | — | — | |
| ISyNet-N12021.09 | 97.92 | — | — | — | — | — | — | |
| ResNet-152# Params=58.1M, # FLOPs=11.3B, Pre-training=ImageNet, Fine-tuning=true2021.07 | 97.9 | — | — | — | — | — | — | |
| Inception-v4# Params=41.1M, # FLOPs=16.1B, Pre-training=ImageNet, Fine-tuning=true2021.07 | 97.9 | — | — | — | — | — | — | |
| ISyNet-N1-S32021.09 | 97.88 | — | — | — | — | — | — | |
| ISyNet-N1-S12021.09 | 97.86 | — | — | — | — | — | — | |
| ViT-B/32#Params=86M2022.04 | 97.8 | — | — | — | — | — | — | |
| ISyNet-N02021.09 | 97.55 | — | — | — | — | — | — | |
| Direct FTMethod Category=PT, Backbone=DeiT-S, 11.3M, Para. (M)=11.02024.06 | 97.5 | — | — | — | — | — | — | |
| WAVEMethod Category=Learngene, Backbone=DeiT-S, 11.3M, Para. (M)=4.02024.06 | 97.4 | — | — | — | — | — | — | |
| ResNet-18+2021.09 | 97.35 | — | — | — | — | — | — | |
| OpenCLIP VIT-H/14zero-shot=true2023.03 | 97.3 | — | — | — | — | — | — | |
| Auto-LGMethod Category=Learngene, Backbone=DeiT-S, 11.3M, Para. (M)=7.52024.06 | 97.3 | — | — | — | — | — | — | |
| TLEGMethod Category=Learngene, Backbone=DeiT-S, 11.3M, Para. (M)=3.92024.06 | 97.2 | — | — | — | — | — | — | |
| Negative Gradient2023.03 | 97.16 | — | — | — | — | — | — | |
| SLFBackbone=WRN2023.10 | 97.1 | — | — | — | — | — | — | |
| LIGOMethod Category=Trans., Backbone=DeiT-S, 11.3M, Para. (M)=7.52024.06 | 96.9 | — | — | — | — | — | — | |
| ALABackbone=WRN2023.10 | 96.7 | — | — | — | — | — | — | |
| L2T-DLNBackbone=WRN2023.10 | 96.7 | — | — | — | — | — | — | |
| L2T-DLFBackbone=WRN2023.10 | 96.6 | — | — | — | — | — | — | |
| MixOEBackbone=ResNet-182024.05 | 96.6 | — | — | — | — | — | — | |
| WAVEMethod Category=Learngene, Backbone=DeiT-Ti, 3.0M, Para. (M)=1.12024.06 | 96.6 | — | — | — | — | — | — | |
| Direct FTMethod Category=PT, Backbone=DeiT-Ti, 3.0M, Para. (M)=2.92024.06 | 96.6 | — | — | — | — | — | — | |
| Share InitMethod Category=Trans., Backbone=DeiT-S, 11.3M, Para. (M)=2.22024.06 | 96.5 | — | — | — | — | — | — | |
| Auto-LGMethod Category=Learngene, Backbone=DeiT-Ti, 3.0M, Para. (M)=2.02024.06 | 96.4 | — | — | — | — | — | — | |
| L-M SoftmaxBackbone=WRN2023.10 | 96.3 | — | — | — | — | — | — | |
| CEBackbone=WRN2023.10 | 96.2 | — | — | — | — | — | — | |
| SmoothBackbone=WRN2023.10 | 96.2 | — | — | — | — | — | — | |
| TLEGMethod Category=Learngene, Backbone=DeiT-Ti, 3.0M, Para. (M)=1.12024.06 | 96.1 | — | — | — | — | — | — | |
| Share InitMethod Category=Trans., Backbone=DeiT-Ti, 3.0M, Para. (M)=0.62024.06 | 96 | — | — | — | — | — | — | |
| ARLFBackbone=WRN2023.10 | 95.9 | — | — | — | — | — | — | |
| TDDSp (pruning rate)=50%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.66 | — | — | — | — | — | — | |
| LIGOMethod Category=Trans., Backbone=DeiT-Ti, 3.0M, Para. (M)=2.02024.06 | 95.6 | — | — | — | — | — | — | |
| TDDSp (pruning rate)=30%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.5 | — | — | — | — | — | — | |
| EL2Np (pruning rate)=30%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.44 | — | — | — | — | — | — | |
| CCSp (pruning rate)=30%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.4 | — | — | — | — | — | — | |
| Forgettingp (pruning rate)=30%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.36 | — | — | — | — | — | — | |
| Forgettingp (pruning rate)=50%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.29 | — | — | — | — | — | — | |
| AUMp (pruning rate)=50%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.26 | — | — | — | — | — | — | |
| FSVI EnsembleBackbone=ResNet-18, Context Distribution (pxc)=random monochrome2023.12 | 95.19 | — | — | — | — | 1.3 | — | |
| Deep EnsembleBackbone=ResNet-182023.12 | 95.13 | — | — | — | — | 1.9 | — | |
| DALBackbone=ResNet-182024.05 | 95.11 | — | — | — | — | — | — | |
| Heur-LGMethod Category=Learngene, Backbone=DeiT-S, 11.3M, Para. (M)=5.72024.06 | 95.1 | — | — | — | — | — | — | |
| AUMp (pruning rate)=30%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.07 | — | — | — | — | — | — | |
| CCSp (pruning rate)=50%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 95.04 | — | — | — | — | — | — | |
| DOEBackbone=ResNet-182024.05 | 94.93 | — | — | — | — | — | — | |
| FixMatchLearning paradigm=Semi-supervised2021.03 | 94.9 | — | — | — | — | — | — | |
| MSP-OEBackbone=ResNet-182024.05 | 94.83 | — | — | — | — | — | — | |
| DOSBackbone=ResNet-182024.05 | 94.74 | — | — | — | — | — | — | |
| DivOEBackbone=ResNet-182024.05 | 94.72 | — | — | — | — | — | — | |
| ReMixMatchLearning paradigm=Semi-supervised2021.03 | 94.7 | — | — | — | — | — | — | |
| EL2Np (pruning rate)=50%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 94.61 | — | — | — | — | — | — | |
| Entropyp (pruning rate)=30%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 94.44 | — | — | — | — | — | — | |
| Randomp (pruning rate)=30%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 94.33 | — | — | — | — | — | — | |
| DUQBackbone=ResNet-182023.12 | 94.1 | — | — | — | — | — | — | |
| Wt SelectMethod Category=Trans., Backbone=DeiT-S, 11.3M, Para. (M)=11.02024.06 | 94.1 | — | — | — | — | — | — | |
| Hopfield BoostingBackbone=ResNet-182024.05 | 94.02 | — | — | — | — | — | — | |
| Heur-LGMethod Category=Learngene, Backbone=DeiT-Ti, 3.0M, Para. (M)=1.52024.06 | 94 | — | — | — | — | — | — | |
| He-InitMethod Category=Direct, Backbone=DeiT-S, 11.3M, Para. (M)=02024.06 | 94 | — | — | — | — | — | — | |
| TDDSp (pruning rate)=70%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 93.92 | — | — | — | — | — | — | |
| Cold PAWSSelf-supervised=SimCLR, Selection strategy=best k-mediods (t-SNE), Semi-supervised=PAWS, Labels=1002023.05 | 93.9 | — | — | — | — | — | — | |
| GHN-3Method Category=Direct, Backbone=DeiT-S, 11.3M, Para. (M)=02024.06 | 93.9 | — | — | — | — | — | — | |
| Mixed Barlow TwinsBackbone=ResNet-50, Epochs=2000, Evaluation Protocol=linear2023.12 | 93.89 | — | — | — | — | — | — | |
| Moderatep (pruning rate)=30%, Strategy=Strategy-P, Backbone=ResNet-182023.11 | 93.86 | — | — | — | — | — | — | |
| L2T-DLNBackbone=ResNet322023.10 | 93.8 | — | — | — | — | — | — | |
| Cold PAWSSelf-supervised=SimCLR, Selection strategy=best k-mediods (t-SNE), Semi-supervised=PAWS, Labels=302023.05 | 93.6 | — | — | — | — | — | — | |
| SLFBackbone=ResNet322023.10 | 93.6 | — | — | — | — | — | — | |
| FSVIBackbone=ResNet-18, Context Distribution (pxc)=CIFAR-1002023.12 | 93.57 | — | — | — | — | 2.6 | — | |
| MC DROPOUTBackbone=ResNet-182023.12 | 93.55 | — | — | — | — | 4 | — | |
| Wang et al. (USL-T)Self-supervised=SimCLR, Selection strategy=USL-T, Semi-supervised=FixMatch, Labels=402023.05 | 93.5 | — | — | — | — | — | — | |
| Mixed Barlow TwinsBackbone=ResNet-50, Epochs=1000, Evaluation Protocol=linear2023.12 | 93.48 | — | — | — | — | — | — |