/keras-alchemy

keras alchemy trick

Primary LanguagePython

keras-alchemy

keras alchemy of image classification

Model list(net args)

MobileNet
MobileNetV2
MobileNetV3Small
MobileNetV3Large
NASNetMobile
NASNetLarge
SqueezeNet
ShuffleNet
ShuffleNetV2
MixNetSmall
MixNetMedium
MixNetLarge
ResNet18
ResNet34
ResNet50
ResNet101
ResNet152
ResNet50V2
ResNeXt50
DenseNet21
DenseNet121
AttentionResNet56
AttentionResNet92
DPN92
DPN98
DPN107
DPN137
DarkNet53
MnasNet
SimpleNetV1
SimpleNetV2
VGG11
VGG13
VGG16
VGG19
VGG19Small
EfficientNetB0
ResDenseNet-30
MultiResNet
BlazeFace

cifar100 Results

network params val_acc
SimpleNetV1 18,113,508 0.7295
SimpleNetV2 5,522,739 0.7305
SimpleNetV2Capsule 6,861,839 0.7583
SqueezeNet 787,108 0.7068
SqueezeCapsule 2,374,208 0.7295
ShuffleNet 1,027,324 0.6708
ShuffleNetV2 4,121,336 0.7205
ShuffleNetV2Capsule 7,295,636 0.7125
MobileNet 3,331,364 0.7305
MobileNetV2 2,386,084 0.6496
MobileNetV3Small 3,073,916 0.6503
MobileNetV3Large 5,252,092 0.7387
VGG11 28,523,748 0.6621
VGG11Small 2,932,996 0.6330
VGG19 39,338,660 0.6791
VGG19Small 5,114,852 0.6578
ResNet18 11,237,156 0.7563
ResNet34 21,352,740 0.7785
ResNet50V2 23,769,700 0.7593
ResDenseNet-30 1,524,724 0.7305
ResDenseNet-30-Capsule 3,719,424 0.7402
ResDenseNet-53 2,976,004 0.7559
ResDenseNet-104 7,407,972 0.7689
MnasNet 3,217,900 0.7521
MultiResNet 6,014,038 0.6921
BlazeFace 715,044 0.6236

Reference