List your company
ResearchFormerly WizWand

Image Classification on Food

95.3AccuracyDom-Ad (Am-B)
69.09275.89682.789.504Sep 28, 2020Sep 6, 2021Aug 15, 2022Jul 24, 2023Jul 1, 2024Jun 9, 2025May 18, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Dom-Ad (Am-B)Backbone Architecture=AmoebaNet-B2020.0995.3——
Dom-Ad (In-v3)Backbone Architecture=Inception-v32020.0994.1——
KNNSelection Method=KNN, Adaptation Strategy=Full, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.0993.1——
KLSelection Method=KL, Adaptation Strategy=Full, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.0993.1——
KNNSelection Method=KNN, Adaptation Strategy=Full, Training Epochs=2e, Backbone Architecture=Resnet-50-v22020.0992.9——
KLSelection Method=KL, Adaptation Strategy=Full, Training Epochs=2e, Backbone Architecture=Resnet-50-v22020.0992.9——
OpenAI-CLIP ViT-LN=400M, Zero-shot=true2024.0992.9——
OpenAI-CLIP VIT-LN=400M, Zero-shot=true2024.0992.9——
EPNSelection Method=EPN, Adaptation Strategy=Full, Training Epochs=2e, Backbone Architecture=Resnet-50-v22020.0992.5——
EPNSelection Method=EPN, Adaptation Strategy=Full, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.0992.4——
OriginalBackbone=ViT-L/14, Step=Step 02026.0592.05——
FTBackbone=ViT-L/14, Step=Step 12026.0591.72——
KNNSelection Method=KNN, Adaptation Strategy=Adapters, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.0991.6——
KLSelection Method=KL, Adaptation Strategy=Adapters, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.0991.6——
EPNSelection Method=EPN, Adaptation Strategy=Adapters, Training Epochs=4e, Backbone Architecture=Resnet-50-v22020.0991.5——
GABackbone=ViT-L/14, Step=Step 12026.0591.5——
X+OS_SimCoreCriterion=Stopping Criterion2023.0391.31——
BaselineBackbone Architecture=Resnet-50-v22020.0991.3——
CATABackbone=ViT-L/14, Step=Step 22026.0591.23——
CATABackbone=ViT-L/14, Step=Step 12026.0591.22——
CATABackbone=ViT-L/14, Step=Step 32026.0591.13——
CATABackbone=ViT-L/14, Step=Step 52026.0591.11——
CATABackbone=ViT-L/14, Step=Step 42026.0590.99——
FTBackbone=ViT-L/14, Step=Step 32026.0590.52——
FTBackbone=ViT-L/14, Step=Step 22026.0590.36——
FTBackbone=ViT-L/14, Step=Step 42026.0590.08——
LAION-CLIP ViT-LN=400M, Zero-shot=true2024.0990——
LAION-CLIP VIT-LN=400M, Zero-shot=true2024.0990——
TWISTEvaluation Protocol=Fine-tune, Backbone=ResNet-50, Pre-trained=ImageNet2021.1089.3——
FTBackbone=ViT-L/14, Step=Step 52026.0589.23——
DINOv2-ARLN=20M, Zero-shot=true2024.0989.1——
DINOv2-ARLN=20M, Zero-shot=true2024.0989.1——
GABackbone=ViT-L/14, Step=Step 22026.0588.91——
BYOLEvaluation Protocol=Fine-tune, Backbone=ResNet-50, Pre-trained=ImageNet2021.1088.5——
BYOLEvaluation Protocol=Fine-tuned, Backbone=ResNet, Pre-training Dataset=ImageNet2020.1188.5——
SUPEvaluation Protocol=Fine-tune, Backbone=ResNet-50, Pre-trained=ImageNet2021.1088.3——
FNCEvaluation Protocol=Fine-tuned, Backbone=ResNet, Pre-training Dataset=ImageNet2020.1188.3——
SimCLR v1Evaluation Protocol=Fine-tuned, Backbone=ResNet, Pre-training Dataset=ImageNet2020.1188.2——
SimCLR v2Evaluation Protocol=Fine-tuned, Backbone=ResNet, Pre-training Dataset=ImageNet2020.1188.2——
InMaPBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.0488——
X2023.0387.99——
InMaP + ZLaPBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.0487.9—-0.1
TransCLIP-FSShots=1, Backbone=ViT-B/162024.0687.9——
SimCLREvaluation Protocol=Fine-tune, Backbone=ResNet-50, Pre-trained=ImageNet2021.1087.5——
DINOv2-MpNetN=20M, Zero-shot=true2024.0987.5——
DINOv2-MpNetN=20M, Zero-shot=true2024.0987.5——
CLIP + ZLaPBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.0487.2—1.5
RandomEvaluation Protocol=Fine-tune, Backbone=ResNet-50, Pre-trained=ImageNet2021.1086.9——
TransCLIP-FSShots=16, Backbone=ViT-B/162024.0686.9——
TransCLIP-FSShots=4, Backbone=ViT-B/162024.0686.4——
CLIP-DNBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.0486——
CLIP-ViT-B/16Shots=0, Backbone=ViT-B/162024.0685.9——
CLIPBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.0485.7——
TPTBackbone=ViT-B/16, Evaluation Protocol=Transductive Zero-Shot2024.0484.7——
REACTEvaluation Protocol=Retrieval Augmented, Retrieval Size=10K, Variant=Gated-Image2024.0183——
FTBackbone=ViT-B/32, Step=Step 12026.0582.17——
OriginalBackbone=ViT-B/32, Step=Step 02026.0582.05——
REACTEvaluation Protocol=Retrieval Augmented, Retrieval Size=10K, Variant=Locked-Text2024.0181.8——
GABackbone=ViT-B/32, Step=Step 12026.0581.8——
REAL-LinearEvaluation Protocol=Retrieval Augmented, Retrieval Size=5002024.0181.4——
FTBackbone=ViT-B/32, Step=Step 22026.0581.22——
LiT L16LN=112M, Zero-shot=true2024.0981.1——
LiT L16LN=112M, Zero-shot=true2024.0981.1——
OriginalBackbone=RN1012026.0581.1——
REAL-PromptEvaluation Protocol=Zero-Shot Prompting2024.0181——
OpenAI templatesEvaluation Protocol=Zero-Shot Prompting2024.0180.9——
PromptingEvaluation Protocol=Zero-Shot Prompting, Prompt Template=a photo of {concept}2024.0180.3——
PromptingEvaluation Protocol=Zero-Shot Prompting, Prompt Template={concept}2024.0180.1——
DCLIPEvaluation Protocol=Zero-Shot Prompting2024.0180.1——
FTBackbone=ViT-B/32, Step=Step 32026.0579.93——
CuPLEvaluation Protocol=Zero-Shot Prompting2024.0179.5——
InMaPBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.0479.4——
GABackbone=ViT-B/32, Step=Step 22026.0579.3——
InMaP + ZLaPBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.0479.2—-0.2
CLIP-LIPBackbone=RN1012026.0579.11——
CATABackbone=RN1012026.0578.97——
CLIP + ZLaPBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.0478.8—2.2
FTBackbone=ViT-B/32, Step=Step 42026.0578.72——
CATABackbone=ViT-B/32, Step=Step 22026.0578.42——
CATABackbone=ViT-B/32, Step=Step 12026.0578.18——
CATABackbone=ViT-B/32, Step=Step 32026.0578.06——
CATABackbone=ViT-B/32, Step=Step 42026.0578.04——
FTBackbone=ViT-B/32, Step=Step 52026.0578.02——
TWISTEvaluation Protocol=Linear evaluation, Backbone=ResNet-50, Pre-trained=ImageNet2021.1078——
CLIP-DNBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.0477.2——
CATABackbone=ViT-B/32, Step=Step 52026.0577.2——
GABackbone=ViT-B/32, Step=Step 32026.0577——
CLIP-LIPBackbone=RN502026.0576.78——
CLIPBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.0476.6——
OriginalBackbone=RN502026.0576.49——
BYOLEvaluation Protocol=Linear evaluation, Backbone=ResNet-50, Pre-trained=ImageNet2021.1075.3——
BYOLEvaluation Protocol=Linear eval, Backbone=ResNet, Pre-training Dataset=ImageNet2020.1175.3——
TPTBackbone=ResNet-50, Evaluation Protocol=Transductive Zero-Shot2024.0474.9——
GABackbone=ViT-B/32, Step=Step 42026.0574.47——
CATABackbone=RN502026.0574.43——
FNCEvaluation Protocol=Linear eval, Backbone=ResNet, Pre-training Dataset=ImageNet2020.1174.4——
SimCLR v2Evaluation Protocol=Linear eval, Backbone=ResNet, Pre-training Dataset=ImageNet2020.1173.9——
SUPEvaluation Protocol=Linear evaluation, Backbone=ResNet-50, Pre-trained=ImageNet2021.1072.3——
SimSiam + AugSelfFine-tuning strategy=L2SP, Pre-trained on=ImageNet100, Backbone=ResNet-502021.1170.19——
GABackbone=ViT-B/32, Step=Step 52026.0570.1——