Generate VGG16 model for CoreML

gistfile1.py

# Modified from https://gist.github.com/baraldilorenzo/07d7802847aaad0a35d3#file-vgg-16_keras-py
from keras.models import Sequential
from keras.layers.core import Flatten, Dense, Dropout
from keras.layers.convolutional import Convolution2D, MaxPooling2D, ZeroPadding2D
from keras.optimizers import SGD
import cv2, numpy as np
from keras import backend as K
import csv
import coremltools
from keras import models
import sys
import imghdr
import os.path

def VGG_16(weights_path=None):
    model = Sequential()
    model.add(ZeroPadding2D((1,1),input_shape=(3,224,224)))
    model.add(Convolution2D(64, 3, 3, activation='relu'))
    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(64, 3, 3, activation='relu'))
    model.add(MaxPooling2D((2,2), strides=(2,2)))

    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(128, 3, 3, activation='relu'))
    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(128, 3, 3, activation='relu'))
    model.add(MaxPooling2D((2,2), strides=(2,2)))

    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(256, 3, 3, activation='relu'))
    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(256, 3, 3, activation='relu'))
    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(256, 3, 3, activation='relu'))
    model.add(MaxPooling2D((2,2), strides=(2,2)))

    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(512, 3, 3, activation='relu'))
    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(512, 3, 3, activation='relu'))
    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(512, 3, 3, activation='relu'))
    model.add(MaxPooling2D((2,2), strides=(2,2)))

    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(512, 3, 3, activation='relu'))
    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(512, 3, 3, activation='relu'))
    model.add(ZeroPadding2D((1,1)))
    model.add(Convolution2D(512, 3, 3, activation='relu'))
    model.add(MaxPooling2D((2,2), strides=(2,2)))

    model.add(Flatten())
    model.add(Dense(4096, activation='relu'))
    model.add(Dropout(0.5))
    model.add(Dense(4096, activation='relu'))
    model.add(Dropout(0.5))
    model.add(Dense(1000, activation='softmax'))

    if weights_path:
        model.load_weights(weights_path)

    return model

if __name__ == "__main__":



    K.set_image_dim_ordering('th')

    ## Check if there input filename exists
    if os.path.exists(sys.argv[1]) == False:
        print("Please enter a valid filename")
        exit()

    input_image = cv2.imread(sys.argv[1])
    img = cv2.resize(input_image, (224, 224)).astype(np.float32)
    mean_pixel = [103.939, 116.779, 123.68]
    img = img.astype(np.float32, copy=False)
    for c in range(3):
        img[:, :, c] = img[:, :, c] - mean_pixel[c]
    img = img.transpose((2,0,1))        
    img = np.expand_dims(img, axis=0)

    # Test pretrained model
    model = VGG_16('vgg16_weights.h5')
    sgd = SGD(lr=0.1, decay=1e-6, momentum=0.9, nesterov=True)
    model.compile(optimizer=sgd, loss='categorical_crossentropy')
    out = model.predict(img)
    line =  np.argmax(out)

    the_file = open('synset_words.txt', 'r')
    reader = csv.reader(the_file)
    for i, row in enumerate(reader):
        if i == line:
            print(row)
            break
    models.save_model(model,'vgg16.h5')
    
    ### Uncomment if you want to generate the .mlmodel file for Xcode
    # coreml_model = coremltools.converters.keras.convert('vgg16.h5')
    # coreml_model.save('vgg16_keras.mlmodel')