List your company

Class-Incremental Learning on TinyImageNet (test)

52.77AccuracyNAPA-VQ
17.441226.613135.78544.9569Mar 16, 2023Jul 17, 2023Nov 18, 2023Mar 21, 2024Jul 23, 2024Nov 24, 2024Mar 28, 2025

Evaluation results

MethodLinks
NAPA-VQNumber of incremental tasks (T)=52023.0852.77—————
NAPA-VQNumber of incremental tasks (T)=102023.0851.78—————
SSRENumber of incremental tasks (T)=52023.0850.39—————
PASSNumber of incremental tasks (T)=52023.0849.55—————
NAPA-VQNumber of incremental tasks (T)=202023.0849.51—————
SSRENumber of incremental tasks (T)=102023.0848.93—————
SSRENumber of incremental tasks (T)=202023.0848.17—————
PASSNumber of incremental tasks (T)=102023.0847.29—————
IL2ANumber of incremental tasks (T)=52023.0847.02—————
IL2ANumber of incremental tasks (T)=102023.0844.48—————
Joint2023.0344.01—————
PASSNumber of incremental tasks (T)=202023.0842.07—————
IL2ANumber of incremental tasks (T)=202023.0839.68—————
BiC w/ ANCLframework=ANCL2023.0334.26—————
BiC2023.0332.87—————
PODNet w/ ANCLframework=ANCL2023.0332.53—————
VanillaCalibration Method=None, Backbone=ResNet-18, Base CIL method=ER, Number of incremental tasks=10, Memory size=4000, Batch size=2562025.0331.95—32.5632.55——
TSCalibration Method=Temperature Scaling, Backbone=ResNet-18, Base CIL method=ER, Number of incremental tasks=10, Memory size=4000, Batch size=2562025.0331.48—19.4419.44——
ETSCalibration Method=Ensemble Temperature Scaling, Backbone=ResNet-18, Base CIL method=ER, Number of incremental tasks=10, Memory size=4000, Batch size=2562025.0331.48—19.8119.83——
IRMCalibration Method=Isotonic Regression, Backbone=ResNet-18, Base CIL method=ER, Number of incremental tasks=10, Memory size=4000, Batch size=2562025.0331.48—19.7719.73——
PerturbTSCalibration Method=Perturbed Temperature Scaling, Backbone=ResNet-18, Base CIL method=ER, Number of incremental tasks=10, Memory size=4000, Batch size=2562025.0331.48—10.610.58——
T-CILCalibration Method=T-CIL, Backbone=ResNet-18, Base CIL method=ER, Number of incremental tasks=10, Memory size=4000, Batch size=2562025.0331.48—8.128.12——
PODNet2023.0330.78—————
iCaRL w/ ANCLframework=ANCL2023.0330.36—————
LUCIR w/ ANCLframework=ANCL2023.0328.96—————
iCaRL2023.0327.83—————
LUCIR2023.0324.07—————
EEIL2023.0320.99—————
Fine-tuning2023.0318.8—————
AdaGaussT=10, Training Setting=from scratch2026.06————32.945.8
AdaGaussT=20, Training Setting=from scratch2026.06————27.539.5
ADCT=10, Training Setting=from scratch2026.06————32.343
ADCT=20, Training Setting=from scratch2026.06————18.136
DPCRT=10, Training Setting=from scratch2026.06————34.346.9
DPCRT=20, Training Setting=from scratch2026.06————25.639.3
EEILNumber of phases=52022.03—25.56————
EEILNumber of phases=102022.03—25.91————
EEILNumber of phases=202022.03—35.04————
EFCT=10, Training Setting=from scratch2026.06————34.547.9
EFCT=20, Training Setting=from scratch2026.06————28.442.1
EWCT=10, Training Setting=from scratch2026.06————18.534.3
EWCT=20, Training Setting=from scratch2026.06————11.326.8
FeCAMT=10, Training Setting=from scratch2026.06————30.944.9
FeCAMT=20, Training Setting=from scratch2026.06————24.937.9
FeTrILT=10, Training Setting=from scratch2026.06————3145.3
FeTrILT=20, Training Setting=from scratch2026.06————25.939.9
GATFT=10, Training Setting=from scratch2026.06————37.250.6
GATFT=20, Training Setting=from scratch2026.06————31.545.1
iCaRL-CNNNumber of phases=52022.03—36.89————
iCaRL-CNNNumber of phases=102022.03—36.7————
iCaRL-CNNNumber of phases=202022.03—45.12————
iCaRL-NCMNumber of phases=52022.03—27.15————
iCaRL-NCMNumber of phases=102022.03—28.89————
iCaRL-NCMNumber of phases=202022.03—37.4————
LDCT=10, Training Setting=from scratch2026.06————34.246.8
LDCT=20, Training Setting=from scratch2026.06————24.938.2
LwFT=10, Training Setting=from scratch2026.06————27.139.6
LwFT=20, Training Setting=from scratch2026.06————15.231.5
LwF_MCNumber of phases=52022.03—54.26————
LwF_MCNumber of phases=102022.03—54.37————
LwF_MCNumber of phases=202022.03—63.54————
MUCNumber of phases=52022.03—51.46————
MUCNumber of phases=102022.03—50.21————
MUCNumber of phases=202022.03—58————
PASSNumber of phases=52022.03—18.04————
PASSNumber of phases=102022.03—23.11————
PASSNumber of phases=202022.03—30.55————
PASST=10, Training Setting=from scratch2026.06————24.539.5
PASST=20, Training Setting=from scratch2026.06————18.530.4
SDCT=10, Training Setting=from scratch2026.06————29.543.8
SDCT=20, Training Setting=from scratch2026.06————26.340.6
Self-Sustaining Representation ExpansionNumber of phases=52022.03—9.17————
Self-Sustaining Representation ExpansionNumber of phases=102022.03—14.06————
Self-Sustaining Representation ExpansionNumber of phases=202022.03—14.2————
UCIRNumber of phases=52022.03—20.61————
UCIRNumber of phases=102022.03—22.25————
UCIRNumber of phases=202022.03—33.74————