List your company
ResearchFormerly WizWand

Image Classification on USPS

99.5AccuracyThermoLion
39.75255.263570.77586.2865Dec 1, 2025Dec 25, 2025Jan 19, 2026Feb 13, 2026Mar 9, 2026Apr 3, 2026Apr 28, 2026
Updated 4mo ago

Evaluation results

MethodLinks
ThermoLion2025.1299.5——
BioTuneBackbone=ResNet-502026.0197.570.5—
L1-SPBackbone=ResNet-502026.0197.31——
FTBackbone=ResNet-502026.0197.05——
L2-SPBackbone=ResNet-502026.0197——
LoRABackbone=ResNet-502026.0196.92——
AutoRGNBackbone=ResNet-502026.0196.91——
G-FLBackbone=ResNet-502026.0196.86——
G-LFBackbone=ResNet-502026.0196.72——
Adam2025.1296.64——
MuAdam2025.1295.78——
TeacherBackbone=LeNet52026.0495.47——
DIP-KDBackbone=LeNet5, Synthetic Dataset Size=50K2026.0494.2——
DFHL-RSBackbone=LeNet5, Synthetic Dataset Size=50K2026.0493.97——
IDEALBackbone=LeNet5, Synthetic Dataset Size=50K2026.0492.66——
LPBackbone=ResNet-502026.0191.78——
ZSDB3Backbone=LeNet5, Synthetic Dataset Size=50K2026.0465.5——
NaiveKDBackbone=LeNet5, Synthetic Dataset Size=50K2026.0442.05——
CirculantNumber of random features (k)=2562016.05——7.54
CirculantNumber of random features (k)=12802016.05——4.53
FastfoodNumber of random features (k)=2562016.05——7.37
FastfoodNumber of random features (k)=12802016.05——4.62
GaussianNumber of random features (k)=2562016.05——7.12
GaussianNumber of random features (k)=12802016.05——4.52
QMC (Halton)Number of random features (k)=2562016.05——6.9
QMC (Halton)Number of random features (k)=12802016.05——4.73
ToeplitzLike(1)Number of random features (k)=2562016.05——7.72
ToeplitzLike(1)Number of random features (k)=12802016.05——4.62
ToeplitzLike(10)Number of random features (k)=2562016.05——7.46
ToeplitzLike(10)Number of random features (k)=12802016.05——4.53
ToeplitzLike(20)Number of random features (k)=2562016.05——7.29
ToeplitzLike(20)Number of random features (k)=12802016.05——4.65
ToeplitzLike(5)Number of random features (k)=2562016.05——7.44
ToeplitzLike(5)Number of random features (k)=12802016.05——4.58