Image Classification on Caltech101
98.37Base AccuracyHPT
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| HPT2024.08 | 98.37 | — | 96.65 | — | 94.98 | |
| IVLPShots=162024.04 | 98.28 | 93.65 | 95.91 | — | — | |
| PromptSRC2024.08 | 98.28 | — | 96.39 | — | 94.58 | |
| CoPrompt2024.08 | 98.27 | — | 96.55 | — | 94.9 | |
| HPT++2024.08 | 98.17 | — | 96.96 | — | 95.78 | |
| MetaPrompt2022.12 | 98.13 | 94.58 | 96.32 | — | — | |
| CoOpBackbone=ViT-B/16, Context length (M)=4, Shots=162023.03 | 98.11 | 93.52 | 95.76 | — | — | |
| ProMetaRShots=162024.04 | 98.11 | 94.29 | 96.16 | — | — | |
| PromptSRCShots=162024.04 | 98.1 | 94.03 | 96.02 | — | — | |
| MetaPrompt2024.08 | 98.1 | — | 96.02 | — | 94.03 | |
| UNIGRAMShots=162024.04 | 98.07 | 95.11 | 96.57 | — | — | |
| ProGradBackbone=ViT-B/16, Context length (M)=4, Shots=162023.03 | 98.02 | 93.89 | 95.91 | — | — | |
| ProGrad2024.08 | 98.02 | — | 95.91 | — | 93.89 | |
| CoOp2022.12 | 98 | 89.81 | 93.73 | — | — | |
| CoOpShots=162024.04 | 98 | 89.81 | 93.73 | — | — | |
| CoOp2024.08 | 98 | — | 93.73 | — | 89.81 | |
| RPOShots=162024.04 | 97.97 | 94.37 | 96.03 | — | — | |
| CoCoOpBackbone=ViT-B/16, Context length (M)=4, Shots=162023.03 | 97.96 | 93.81 | 95.84 | — | — | |
| CoCoOp2022.12 | 97.96 | 93.81 | 95.84 | — | — | |
| CoCoOpShots=162024.04 | 97.96 | 93.81 | 95.84 | — | — | |
| CoCoOp2024.08 | 97.96 | — | 95.84 | — | 93.81 | |
| MaPLe2022.12 | 97.74 | 94.36 | 96.02 | — | — | |
| MaPLeShots=162024.04 | 97.74 | 94.36 | 96.02 | — | — | |
| MaPLe2024.08 | 97.74 | — | 96.02 | — | 94.36 | |
| KgCoOpBackbone=ViT-B/16, Context length (M)=4, Shots=162023.03 | 97.72 | 94.39 | 96.03 | — | — | |
| KgCoOp2024.08 | 97.72 | — | 96.03 | — | 94.39 | |
| X-FM_baseLinear evaluation=true, Model size=Base, Patch size=16*16, Resolution=224*2242023.01 | 97.1 | — | — | — | — | |
| CLIPBackbone=ViT-B/162023.03 | 96.84 | 94 | 95.4 | — | — | |
| CLIP2022.12 | 96.84 | 94 | 95.4 | — | — | |
| CLIPShots=Zero-shot2024.04 | 96.84 | 94 | 95.4 | — | — | |
| CLIPmode=zero-shot2024.08 | 96.84 | — | 95.4 | — | 94 | |
| ST-ProtoPNetBackbone=ResNet34, Explanation=Single-Scale2024.04 | 96.24 | — | — | — | — | |
| ST-ProtoPNetBackbone=ResNet152, Explanation=Single-Scale2024.04 | 96.09 | — | — | — | — | |
| Deformable ProtoPNetBackbone=ResNet34, Explanation=Single-Scale2024.04 | 95.95 | — | — | — | — | |
| BaselineBackbone=ResNet152, Explanation=N/A2024.04 | 95.76 | — | — | — | — | |
| FLAVALinear evaluation=true, Model size=Base, Patch size=16*16, Resolution=224*2242023.01 | 95.7 | — | — | — | — | |
| BaselineBackbone=ResNet34, Explanation=N/A2024.04 | 94.78 | — | — | — | — | |
| BASIC-LBackbone=BASIC-L, Evaluation Protocol=Zero-shot2021.11 | 94.7 | — | — | — | — | |
| CLIPLinear evaluation=true, Model size=Base, Patch size=16*16, Resolution=224*2242023.01 | 94.7 | — | — | — | — | |
| UNIGRAM2024.04 | 94.67 | — | — | — | — | |
| Any-shift PromptingEvaluation Protocol=Cross-dataset transfer evaluation2024.02 | 94.57 | — | — | — | — | |
| MCPNetBackbone=ResNet152, Explanation=Multi-Scale2024.04 | 94.54 | — | — | — | — | |
| CoCoOpPrompts=tp+v, Backbone=ViT-16/B, Source Dataset=ImageNet, Shots=162023.03 | 94.43 | — | — | — | — | |
| CoCoOpEvaluation Protocol=Cross-dataset transfer evaluation2024.02 | 94.43 | — | — | — | — | |
| CoCoOp2024.04 | 94.43 | — | — | — | — | |
| Full dataBackbone=ViT-B/32, Evaluation Protocol=Full Data2023.11 | 94.4 | — | — | — | — | |
| BASIC-MBackbone=BASIC-M, Evaluation Protocol=Zero-shot2021.11 | 94.2 | — | — | — | — | |
| VP (white-box)Access=white-box, Shots=16, Repeated runs=32023.03 | 94.2 | — | — | — | — | |
| DaVinciLinear evaluation=true, Model size=Base, Patch size=16*16, Resolution=224*2242023.01 | 94.1 | — | — | — | — | |
| KgCoOpPrompts=tp, Backbone=ViT-16/B, Source Dataset=ImageNet, Shots=162023.03 | 93.92 | — | — | — | — | |
| BADGE + PCB(AS)Backbone=ViT-B/32, Active Learning Strategy=BADGE, Integration Method=PCB(AS), Number of Rounds=82023.11 | 93.85 | — | — | — | — | |
| BARShots=16, Repeated runs=32023.03 | 93.8 | — | — | — | — | |
| ProMetaR2024.04 | 93.74 | — | — | — | — | |
| CoOpPrompts=tp, Backbone=ViT-16/B, Source Dataset=ImageNet, Shots=162023.03 | 93.7 | — | — | — | — | |
| BlackVIPShots=16, Repeated runs=32023.03 | 93.7 | — | — | — | — | |
| CoOpEvaluation Protocol=Cross-dataset transfer evaluation2024.02 | 93.7 | — | — | — | — | |
| CoOp2024.04 | 93.7 | — | — | — | — | |
| BPLEvaluation Protocol=Cross-dataset transfer evaluation2024.02 | 93.67 | — | — | — | — | |
| Deformable ProtoPNetBackbone=ResNet152, Explanation=Single-Scale2024.04 | 93.65 | — | — | — | — | |
| PromptSRC2024.04 | 93.6 | — | — | — | — | |
| MaPLeEvaluation Protocol=Cross-dataset transfer evaluation2024.02 | 93.53 | — | — | — | — | |
| MaPLe2024.04 | 93.53 | — | — | — | — | |
| Entropy + PCB(AS)Backbone=ViT-B/32, Active Learning Strategy=Entropy, Integration Method=PCB(AS), Number of Rounds=82023.11 | 93.48 | — | — | — | — | |
| GraphAdapterSetting=16-shot2023.09 | 93.33 | — | — | — | — | |
| MCPNetBackbone=ResNet34, Explanation=Multi-Scale2024.04 | 93.32 | — | — | — | — | |
| ZSMode=Zero-shot2023.03 | 92.9 | — | — | — | — | |
| TaskResSetting=16-shot2023.09 | 92.9 | — | — | — | — | |
| CLIPBackbone=ViT-L/14-336, Evaluation Protocol=Zero-shot2021.11 | 92.8 | — | — | — | — | |
| Tip-Adapter-FSetting=16-shot2023.09 | 92.63 | — | — | — | — | |
| CLIP-AdapterSetting=16-shot2023.09 | 92.49 | — | — | — | — | |
| GraphAdapterSetting=8-shot2023.09 | 92.45 | — | — | — | — | |
| TaskResSetting=8-shot2023.09 | 92.23 | — | — | — | — | |
| CoOpSetting=16-shot2023.09 | 91.83 | — | — | — | — | |
| Tip-Adapter-FSetting=8-shot2023.09 | 91.7 | — | — | — | — | |
| Coreset + PCB(AS)Backbone=ViT-B/32, Active Learning Strategy=Coreset, Integration Method=PCB(AS), Number of Rounds=82023.11 | 91.7 | — | — | — | — | |
| BASIC-SBackbone=BASIC-S, Evaluation Protocol=Zero-shot2021.11 | 91.6 | — | — | — | — | |
| ProGradPrompts=tp, Backbone=ViT-16/B, Source Dataset=ImageNet, Shots=162023.03 | 91.52 | — | — | — | — | |
| CLIP-AdapterSetting=8-shot2023.09 | 91.4 | — | — | — | — | |
| GraphAdapterSetting=4-shot2023.09 | 90.97 | — | — | — | — | |
| Tip-Adapter-FSetting=4-shot2023.09 | 90.87 | — | — | — | — | |
| TaskResSetting=4-shot2023.09 | 90.63 | — | — | — | — | |
| EVA-02-CLIP-E/14+Zero-shot=true2024.02 | 90.5 | — | — | — | — | |
| CoOpSetting=8-shot2023.09 | 90.21 | — | — | — | — | |
| GraphAdapterSetting=2-shot2023.09 | 90.2 | — | — | — | — | |
| TaskResSetting=2-shot2023.09 | 90.13 | — | — | — | — | |
| CLIP-AdapterSetting=4-shot2023.09 | 89.98 | — | — | — | — | |
| EVA-CLIP-18BZero-shot=true2024.02 | 89.8 | — | — | — | — | |
| Tip-Adapter-FSetting=2-shot2023.09 | 89.61 | — | — | — | — | |
| EVA-CLIP-8BZero-shot=true2024.02 | 89.6 | — | — | — | — | |
| CoOpSetting=4-shot2023.09 | 89.55 | — | — | — | — | |
| RandomBackbone=ViT-B/32, Active Learning Strategy=Random, Number of Rounds=82023.11 | 89.55 | — | — | — | — | |
| EVA-01-CLIP-g/14+Zero-shot=true2024.02 | 89.5 | — | — | — | — | |
| InternVL-CZero-shot=true2024.02 | 89.5 | — | — | — | — | |
| DFN5B-CLIP-H/14Zero-shot=true2024.02 | 89.5 | — | — | — | — | |
| VP w/ SPSA-GCShots=16, Repeated runs=3, Optimizer=SPSA-GC2023.03 | 89.4 | — | — | — | — | |
| OpenCLIP-G/14Zero-shot=true2024.02 | 89.4 | — | — | — | — | |
| CLIP-AdapterSetting=2-shot2023.09 | 89.37 | — | — | — | — | |
| CLIPBackbone=ViT-B/16, Evaluation Protocol=Zero-shot2021.11 | 89.3 | — | — | — | — | |
| GraphAdapterSetting=1-shot2023.09 | 88.9 | — | — | — | — | |
| DFN5B-CLIP-H/14+Zero-shot=true2024.02 | 88.9 | — | — | — | — |