List your company
ResearchFormerly WizWand

Image Classification on UCF101 (Accuracy)

84.9AccuracyTOGA
57.75664.80371.8578.897Jun 16, 2025Aug 19, 2025Oct 22, 2025Dec 26, 2025Feb 28, 2026May 3, 2026Jul 7, 2026
Updated 3mo ago

Evaluation results

MethodLinks
TOGAShots=16, Venue=-2026.0384.9
CLIP-AdapterBackbone=ViT-B/16, Shots=16, Blending ratio (alpha)=best2026.0384
TOGAShots=8, Venue=-2026.0383.9
HOSO-AdapterBackbone=ViT-B/16, Shots=16, Number of runs=32026.0383.43
CLIP-1+DINO(v2+v3)Setting=Inductive, Test-Time Adaptation (TTA)=false2025.0681.7
TOGAShots=4, Venue=-2026.0381.7
IsoCLIPIntra-modal=✓, Classifier=NCM, Backbone=ViT-B/16-open2026.0380.9
CLIP-AdapterBackbone=ViT-B/16, Shots=16, Blending ratio (alpha)=0.22026.0380.4
CLIP-1+DINOv3Setting=Inductive, Test-Time Adaptation (TTA)=false2025.0680.3
Image-ImageIntra-modal=✓, Classifier=NCM, Backbone=ViT-B/16-open2026.0379.9
CLIP-1+DINOv2Setting=Inductive, Test-Time Adaptation (TTA)=false2025.0679.1
TOGAShots=2, Venue=-2026.0377.9
Align+MixShot=16, Backbone=ResNet-502026.0377.9
GDAShot=16, Backbone=ResNet-502026.0376.9
COSMIC*Setting=Inductive, Test-Time Adaptation (TTA)=true2025.0676.2
CLIP-MoE + MoE-GRPOTraining shots=16-shot, Source dataset=ImageNet, Training epochs=32026.0376.2
CLIP-1+CLIP-2Setting=Inductive, Test-Time Adaptation (TTA)=false2025.0675.8
TOGAShots=1, Venue=-2026.0374.5
CLIP-MoE + Det-FTTraining shots=16-shot, Source dataset=ImageNet, Training epochs=32026.0373.9
CLIP-MoETraining shots=16-shot, Source dataset=ImageNet, Training epochs=32026.0373
DMN*Setting=Inductive, Test-Time Adaptation (TTA)=true2025.0672.5
OursTest-time prompt tuning=True2026.0472.24
MaPLe + OursTest-time prompt tuning=True2026.0472.21
TIP-XShot=16, Backbone=ResNet-502026.0372
Align+MixShot=4, Backbone=ResNet-502026.0371.5
TIPPLE*Setting=Inductive, Test-Time Adaptation (TTA)=true2025.0671.2
GDAShot=4, Backbone=ResNet-502026.0371
TDA*Setting=Inductive, Test-Time Adaptation (TTA)=true2025.0670.7
TCP + SAMPLeOptimizer=SAMPLe2026.0770.61
TIP-AdapterShot=16, Backbone=ResNet-502026.0370.6
DPE*Setting=Inductive, Test-Time Adaptation (TTA)=true2025.0670.4
CLIP-SVD2025.0969.91
CoPromptOptimizer=Baseline2026.0769.73
CoPrompt + SAMPLeOptimizer=SAMPLe2026.0769.71
MaPLe + SAMPLeOptimizer=SAMPLe2026.0769.58
ZeroSetting=Inductive, Test-Time Adaptation (TTA)=false2025.0669.2
CoCoOp + SAMPLeOptimizer=SAMPLe2026.0769
TCPOptimizer=Baseline2026.0768.73
MTA*Setting=Inductive, Test-Time Adaptation (TTA)=true2025.0668.7
MaPLeTraining shots=16-shot, Source dataset=ImageNet, Training epochs=32026.0368.7
MaPLe2025.0968.69
MaPLeOptimizer=Baseline2026.0768.69
KgCoOp2025.0968.5
Co-CoOp2025.0968.21
CoCoOpOptimizer=Baseline2026.0768.21
CoCoOpTraining shots=16-shot, Source dataset=ImageNet, Training epochs=32026.0368.2
Image-TextIntra-modal=✗, Classifier=Zero-Shot, Backbone=ViT-B/16-open2026.0367.8
CoOp + SAMPLeOptimizer=SAMPLe2026.0767.64
CLIP-1Setting=Inductive, Test-Time Adaptation (TTA)=false2025.0667.5
TIP-XShot=4, Backbone=ResNet-502026.0366.9
CLIP2025.0966.83
CoOpTraining shots=16-shot, Source dataset=ImageNet, Training epochs=32026.0366.6
CoOp2025.0966.55
CoOpOptimizer=Baseline2026.0766.55
TIP-AdapterShot=4, Backbone=ResNet-502026.0366.5
Align+MixShot=1, Backbone=ResNet-502026.0364.3
ProGrad2025.0964.29
TIP-XShot=1, Backbone=ResNet-502026.0362.8
TIP-AdapterShot=1, Backbone=ResNet-502026.0362.6
CALIPShot=0, Backbone=ResNet-502026.0361.7
CLIPShots=0, Venue=ICML’222026.0361.5
CLIPShot=0, Backbone=ResNet-502026.0361.5
GDAShot=1, Backbone=ResNet-502026.0361.2
CLIP-2Setting=Inductive, Test-Time Adaptation (TTA)=false2025.0658.8