Skip to content

Latest commit

 

History

History
29 lines (23 loc) · 1.41 KB

README.md

File metadata and controls

29 lines (23 loc) · 1.41 KB

KEKOxMNIST

Classification & generative models on MNIST, implemented by Keras.

Classification models

Units: accuracy %

Model Validation Test Comment
Simple MLP 0.0% 0.0%
Simple convnet 0.0% 0.0%
VGG-like convnet 0.0% 0.0%
VGG16 99.61% 99.68% Batch size: 64, Epoch: 200, Image standardization, Data augmentation: rotating(15), width/height shift(0.1), shearing(0.2), zooming(0.1)
Mobilenet 99.63% 99.68% Batch size: 64, Epoch: 200, Image standardization, Data augmentation: rotating(15), width/height shift(0.1), shearing(0.2), zooming(0.1)
Resnet164 99.72% 99.70% Batch size: 128, Epoch: 200, Image standardization, Data augmentation: rotating(15), width/height shift(0.1), shearing(0.2), zooming(0.1)
WideResnet28-10 99.72% 99.76% Batch size: 128, Epoch: 200, Image standardization, Data augmentation: rotating(15), width/height shift(0.1), shearing(0.2), zooming(0.1)

Generative models

Model Sample Comment
GAN
DCGAN
cGAN