Adaptation of VGG net for Keras, with 128x128 greyscale images and 196 target classes

vgg_net.py

import os 
import sys
import json
import model_control 
from keras.models import Sequential
from keras.layers import Dense, Dropout, Activation, Flatten
from keras.layers import Convolution2D, MaxPooling2D
from keras.optimizers import SGD
import numpy as np
from numpy import loadtxt, asarray
from scipy.ndimage import imread

Y_train = loadtxt(model_control.y_train_file, delimiter=',', dtype = int)
train_files = os.listdir(model_control.train_img_path)
train_files = ['%s/%s' % (model_control.train_img_path, x) for x in train_files if 'jpg' in x]
X_train = asarray([imread(x) for x in train_files])
X_train = X_train.reshape(X_train.shape[0], 1, 128, 128)

model = Sequential()
model.add(Convolution2D(64, 3, 3, border_mode='valid', input_shape=(1, 128, 128)))
model.add(Activation('relu'))
model.add(Convolution2D(64, 3, 3))
model.add(Activation('relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(Dropout(0.25))

model.add(Convolution2D(128, 3, 3))
model.add(Activation('relu'))
model.add(Convolution2D(128, 3, 3))
model.add(Activation('relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(Dropout(0.25))

model.add(Convolution2D(256, 3, 3))
model.add(Activation('relu'))
model.add(Convolution2D(256, 3, 3))
model.add(Activation('relu'))
model.add(Convolution2D(256, 3, 3))
model.add(Activation('relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(Dropout(0.25))

model.add(Convolution2D(512, 3, 3))
model.add(Activation('relu'))
model.add(Convolution2D(512, 3, 3))
model.add(Activation('relu'))
model.add(Convolution2D(512, 3, 3))
model.add(Activation('relu'))
model.add(MaxPooling2D((2,2), strides=(2,2)))
model.add(Dropout(0.25))

model.add(Flatten())
model.add(Dense(2000))
model.add(Activation('relu'))
model.add(Dropout(0.5))
model.add(Dense(2000))
model.add(Activation('relu'))
model.add(Dropout(0.5))
model.add(Dense(196, activation='softmax'))

sgd = SGD(lr=model_control.l_rate, decay=0, momentum=0.9, nesterov=True)
model.compile(loss='categorical_crossentropy', optimizer=sgd)

history = model.fit(X_train, Y_train, batch_size=model_control.batch_size, nb_epoch=model_control.nb_epoch, verbose=1)

model.save_weights('vgg_net_weights.h5')
json_string = model.to_json()
with open('vgg_net_structure.json', 'wb') as outfile:
    json.dump(json_string, outfile)