search

PacktPublishing / Neural-Networks-with-Keras-Cookbook

Neural Networks with Keras Cookbook, published by Packt
MIT License
50 stars 49 forks source link

Chapter14/Handwritten_text_recognition discards 6 first layers #3

Open johngrabner opened 4 years ago

johngrabner commented 4 years ago

The line inner = Conv2D(128, (3,3), padding='same')(input_data)

appears to wipe out the 6 layers above. Should this be inner = Conv2D(128, (3,3), padding='same')(inner)

johngrabner commented 4 years ago

if so, then need to drop some of the subsequent MaxPooling2D. if not, then these serve no purpose

inner = Conv2D(32, (3,3), padding='same')(input_data)
inner = Activation('relu')(inner)
inner = MaxPooling2D(pool_size=(2,2),name='max1')(inner)

inner = Conv2D(64, (3,3), padding='same')(inner)
inner = Activation('relu')(inner)
inner = MaxPooling2D(pool_size=(2,2),name='max2')(inner)