List your company
ResearchFormerly WizWand

Image Classification on Caltech-101

98.6Top-1 AccuracyPromptAlign
82.4886.66590.8595.035Nov 22, 2021Jul 26, 2022Mar 29, 2023Dec 1, 2023Aug 3, 2024Apr 6, 2025Dec 9, 2025
Updated 8mo ago

Evaluation results

MethodLinks
PromptAlign2023.1198.6——
MaPLe+TPTtest-time prompt tuning=true2023.1198.54——
ProDA2023.1198.27——
MaPLe2023.1198——
LOUPEevaluation=linear probing2022.0897.5——
Florence-CoSwin-HResolution=384px, Backbone=CoSwin-H, Evaluation Protocol=Linear Probing2021.1196.6——
CLIPevaluation=linear probing2022.0896.5——
PEFTShots=102023.0596.1——
CLIP-ViT-L/14Resolution=336px, Evaluation Protocol=Linear Probing2021.1196——
OURSShots=8, Backbone=ViT-B/16, Templates=72025.1296——
OURSShots=8, Backbone=ViT-B/16, Templates=12025.1295.9——
CLIP-LoRAShots=8, Backbone=ViT-B/16, Templates=72025.1295.8——
OURSShots=4, Backbone=ViT-B/16, Templates=12025.1295.7——
CLIP-LoRAShots=8, Backbone=ViT-B/16, Templates=12025.1295.6——
OURSShots=4, Backbone=ViT-B/16, Templates=72025.1295.5——
CLIP-ResNet-50x64Evaluation Protocol=Linear Probing2021.1195.4——
LoRA-CLIPShots=4, Backbone=ViT-B/16, Templates=12025.1295.2——
LoRA-CLIPShots=4, Backbone=ViT-B/16, Templates=72025.1295.1——
SimCLRv2Backbone=ResNet-152x3, Evaluation Protocol=Linear Probing2021.1194.9——
OURSShots=1, Backbone=ViT-B/16, Templates=12025.1294.8——
ViT-L/16Resolution=384px, Evaluation Protocol=Linear Probing2021.1194.7——
EfficientNet-L2Resolution=800px, Evaluation Protocol=Linear Probing2021.1194.7——
LlipData Size=2.5B, Vision Encoder=ViT-B/16, Evaluation Protocol=Zero-shot2024.1294.7——
CoOpShots=102023.0594.6——
PEFTShots=52023.0594.5——
OURSShots=1, Backbone=ViT-B/16, Templates=72025.1294.5——
CaCoBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0394.4——
BYOLBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0394.2——
LoRA-CLIPShots=1, Backbone=ViT-B/16, Templates=72025.1294.1——
LOUPEmode=zero-shot2022.0893.9——
LoRA-CLIPShots=1, Backbone=ViT-B/16, Templates=12025.1293.7——
LaFTerBackbone=ViT-B/322023.0593.3——
LaFTerShots=02023.0593.3——
MetaCLIPData Size=2.5B, Vision Encoder=ViT-B/16, Evaluation Protocol=Zero-shot2024.1293.3——
CoOpShots=52023.0593.2——
OpenCLIPData Size=2B, Vision Encoder=ViT-B/16, Evaluation Protocol=Zero-shot2024.1293.2——
LlipPre-training Data=2.5B, Model Architecture=ViT-B/32, Zero-shot=true2024.1292.9——
MetaCLIPPre-training Data=2.5B, Model Architecture=ViT-B/32, Zero-shot=true2024.1292.8——
CLIPmode=zero-shot2022.0892.6——
WukongBackbone=ViT-L, Zero-shot=true, Pre-training Dataset=Wukong2022.0292.4——
CLIPBackbone=ViT-L, Zero-shot=true, Pre-training Dataset=Wukong2022.0291.9——
AdCo +Backbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0391.8——
CoOpShots=12023.0591.7——
FILIPBackbone=Swin-L, Zero-shot=true, Pre-training Dataset=Wukong2022.0291.6——
xCLIPLinear probing=true, Backbone=ViT-B/16, Pre-train=IT35M2022.1091.6——
OpenCLIPPre-training Data=DataComp-1B, Model Architecture=ViT-B/32, Zero-shot=true2024.1291.6——
MoCo v2+Backbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0391.4——
NNCLRBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0391.3——
DebiasMatchPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0291——
CLIPBackbone=Swin-L, Zero-shot=true, Pre-training Dataset=Wukong2022.0290.7——
DST (FlexMatch)Pre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0290.6——
UPLBackbone=ViT-B/322023.0590.6——
PEFTShots=12023.0590.6——
nCLIPLinear probing=true, Backbone=ViT-B/16, Pre-train=IT35M2022.1090.5——
CLIPBackbone=ViT-B/322023.0590.5——
OpenCLIPPre-training Data=LAION-2B, Model Architecture=ViT-B/32, Zero-shot=true2024.1290.5——
DST (FlexMatch)Pre-training=Unsupervised, Backbone=ResNet-50, Labels per category=42022.0290.4——
DST (FixMatch)Pre-training=Unsupervised, Backbone=ResNet-50, Labels per category=42022.0290.1——
CLIPLinear probing=true, Backbone=ViT-B/16, Pre-train=IT35M2022.1090——
FILIPBackbone=ViT-L, Zero-shot=true, Pre-training Dataset=Wukong2022.0289.9——
WukongBackbone=Swin-L, Zero-shot=true, Pre-training Dataset=Wukong2022.0289.8——
DST (FixMatch)Pre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0289.6——
SimCLRBackbone=ResNet-50, Pre-training Dataset=ImageNet-1K, Pre-training Epochs=8002022.0389.3——
CLIPMode=Zero-shot2024.0489.3——
CLIPBackbone=ViT-B, Zero-shot=true, Pre-training Dataset=Wukong2022.0289.2——
WukongBackbone=ViT-B, Zero-shot=true, Pre-training Dataset=Wukong2022.0289.1——
Self-TuningPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0288.6——
ARFFinetuning Dataset=ImageNet2024.0488.6——
OpenCLIPPre-training Data=LAION-400M, Model Architecture=ViT-B/32, Zero-shot=true2024.1288.4——
DreamLIPData Size=30M, Vision Encoder=ViT-B/16, Evaluation Protocol=Zero-shot2024.1288——
FLYPFinetuning Dataset=ImageNet2024.0487.6——
FlexMatchPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0287.1——
FLAIRData Size=30M, Vision Encoder=ViT-B/16, Evaluation Protocol=Zero-shot2024.1286.5——
FlexMatchPre-training=Unsupervised, Backbone=ResNet-50, Labels per category=42022.0286.4——
DebiasMatchPre-training=Unsupervised, Backbone=ResNet-50, Labels per category=42022.0286.4——
Pseudo LabelPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0286.3——
FixMatchPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0286.3——
Pseudo LabelPre-training=Unsupervised, Backbone=ResNet-50, Labels per category=42022.0286.2——
DreamLIPPre-training Data=Merged-30M, Model Architecture=ViT-B/32, Zero-shot=true2024.1286.1——
COSMOSPre-training Data=Merged-30M, Model Architecture=ViT-B/32, Zero-shot=true2024.1286.1——
UDAPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0285.8——
COSMOSPre-training Data=CC12M, Model Architecture=ViT-B/32, Zero-shot=true2024.1285.5——
MixMatchPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0285.4——
UDAPre-training=Unsupervised, Backbone=ResNet-50, Labels per category=42022.0285——
CLIP-PRBackbone=ViT-B/322023.0584.8——
OURS_SepProtocol=Linear probing2023.0384.45——
SigLIPPre-training Data=Merged-30M, Model Architecture=ViT-B/32, Zero-shot=true2024.1284.4——
SigLIPData Size=30M, Vision Encoder=ViT-B/16, Evaluation Protocol=Zero-shot2024.1284.3——
VATPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0284.1——
MixMatchPre-training=Unsupervised, Backbone=ResNet-50, Labels per category=42022.0284.1——
RATPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0284——
LP-FTFinetuning Dataset=ImageNet2024.0484——
CLIPPre-training Data=Merged-30M, Model Architecture=ViT-B/32, Zero-shot=true2024.1283.8——
Mean TeacherPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0283.7——
CLIPData Size=30M, Vision Encoder=ViT-B/16, Evaluation Protocol=Zero-shot2024.1283.7——
PI-ModelPre-training=Supervised, Backbone=ResNet-50, Labels per category=42022.0283.5——
OURS_GCProtocol=Linear probing2023.0383.23——
DreamLIPPre-training Data=CC12M, Model Architecture=ViT-B/32, Zero-shot=true2024.1283.2——
FILIPBackbone=ViT-B, Zero-shot=true, Pre-training Dataset=Wukong2022.0283.1——
FixMatchPre-training=Unsupervised, Backbone=ResNet-50, Labels per category=42022.0283.1——