List your company
ResearchFormerly WizWand

Image Classification on EMNIST Letters (test)

95.96AccuracyWaveMix
16.29636.97857.6678.342Jul 7, 2020Dec 14, 2020May 24, 2021Oct 31, 2021Apr 10, 2022Sep 17, 2022Feb 25, 2023
Updated 7mo ago

Evaluation results

MethodLinks
WaveMixAugmentation=TrivialAugment2022.0595.96——
Previous SOTA [73]2022.0595.88——
VGG-5 (Spinal FC)Size of Fully Connected Layer=4HL, 128 Neurons Per Layer, Epoch=200, Parameters=3.630M2020.0795.7995.880.5
VGG-5Size of Fully Connected Layer=1HL, 512 Neurons, Epoch=200, Parameters=3.646M2020.0795.7195.86—
CNN (Spinal FC)Size of Fully Connected Layer=6HL, 10 Neurons Per Layer, Epoch=8, Parameters=16.05k2020.0790.0290.2321.4
CNN (Spinal FC)Size of Fully Connected Layer=6HL, 8 Neurons Per Layer, Epoch=8, Parameters=13.82k2020.0789.8890.0720.11
CNNSize of Fully Connected Layer=1HL, 50 Neurons, Epoch=8, Parameters=21.84k2020.0787.2987.57—
FedCILLocal training iteration (T)=400, Global communication round (R)=200, Mini-batch size (B)=32, Backbone=3-layer CNN, Optimizer=Adam, Learning rate=1e-42023.0278.15——
FedProx+ACGAN ReplayLocal training iteration (T)=400, Global communication round (R)=200, Mini-batch size (B)=32, Backbone=3-layer CNN, Optimizer=Adam, Learning rate=1e-42023.0273.91——
FedAvg+ACGAN ReplayLocal training iteration (T)=400, Global communication round (R)=200, Mini-batch size (B)=32, Backbone=3-layer CNN, Optimizer=Adam, Learning rate=1e-42023.0273.85——
FedAvg+DGRLocal training iteration (T)=400, Global communication round (R)=200, Mini-batch size (B)=32, Backbone=3-layer CNN, Optimizer=Adam, Learning rate=1e-42023.0271.92——
FedProx+DGRLocal training iteration (T)=400, Global communication round (R)=200, Mini-batch size (B)=32, Backbone=3-layer CNN, Optimizer=Adam, Learning rate=1e-42023.0271.83——
FedLwF-2TLocal training iteration (T)=400, Global communication round (R)=200, Mini-batch size (B)=32, Backbone=3-layer CNN, Optimizer=Adam, Learning rate=1e-42023.0223.91——
FedProxLocal training iteration (T)=400, Global communication round (R)=200, Mini-batch size (B)=32, Backbone=3-layer CNN, Optimizer=Adam, Learning rate=1e-42023.0219.69——
FedAvgLocal training iteration (T)=400, Global communication round (R)=200, Mini-batch size (B)=32, Backbone=3-layer CNN, Optimizer=Adam, Learning rate=1e-42023.0219.36——