將 keras 模型導出到 tflite

Question

我正在嘗試結合這兩個示例並為我的 android 應用程序創建 tflite 文件。

https://medium.com/nybles/create-your-first-image-recognition-classifier-using-cnn-keras-and-tensorflow-backend-6eaab98d14dd

https://medium.com/@xianbao.qian/convert-keras-model-to-tflite-e2bdf28ee2d2

這是我的代碼：

# Part 1 - Building the CNN

# Importing the Keras libraries and packages
from keras.models import Sequential
from keras.layers import Convolution2D
from keras.layers import MaxPooling2D
from keras.layers import Flatten
from keras.layers import Dense
import tensorflow as tf
from keras.models import load_model


# Initialising the CNN
classifier = Sequential()

# Step 1 - Convolution
classifier.add(Convolution2D(32, 3, 3, input_shape = (64, 64, 3), activation = 'relu'))

# Step 2 - Pooling
classifier.add(MaxPooling2D(pool_size = (2, 2)))

# Adding a second convolutional layer
classifier.add(Convolution2D(32, 3, 3, activation = 'relu'))
classifier.add(MaxPooling2D(pool_size = (2, 2)))

# Step 3 - Flattening
classifier.add(Flatten())

# Step 4 - Full connection
classifier.add(Dense(output_dim = 128, activation = 'relu'))
classifier.add(Dense(output_dim = 1, activation = 'sigmoid'))

# Compiling the CNN
classifier.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])

# Part 2 - Fitting the CNN to the images

from keras.preprocessing.image import ImageDataGenerator

train_datagen = ImageDataGenerator(rescale = 1./255,
                                   shear_range = 0.2,
                                   zoom_range = 0.2,
                                   horizontal_flip = True)

test_datagen = ImageDataGenerator(rescale = 1./255)

training_set = train_datagen.flow_from_directory('dataset/training_set',
                                                 target_size = (64, 64),
                                                 batch_size = 32,
                                                 class_mode = 'binary')

test_set = test_datagen.flow_from_directory('dataset/test_set',
                                            target_size = (64, 64),
                                            batch_size = 32,
                                            class_mode = 'binary')

classifier.fit_generator(training_set,
                         samples_per_epoch = 80,
                         nb_epoch = 1,
                         validation_data = test_set,
                         nb_val_samples = 20)



output_names = [node.op.name for node in classifier.outputs]
sess = tf.keras.backend.get_session()
frozen_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, output_names)    


tflite_model = tf.contrib.lite.toco_convert(frozen_def, [inputs], output_names)
with tf.gfile.GFile(tflite_graph, 'wb') as f:
    f.write(tflite_model)                     

在這一行：

frozen_def = tf.graph_util.convert_variables_to_constants(sess, sess.graph_def, output_names)

我有一個例外：

tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value conv2d_1/bias
[[Node: _retval_conv2d_1/bias_0_0 = _Retval[T=DT_FLOAT, index=0, _device="/job:localhost/replica:0/task:0/device:CPU:0"](conv2d_1/bias)]]

我是機器學習的初學者，完全不知道這個錯誤是什么:-(

有人可以向我解釋什么是錯的嗎？ 我所需要的只是處理多個包含許多圖片的文件夾，並可以預測新圖片與特定文件夾的關系。 謝謝你。

Answer 1

可以使用tf.lite.TFLiteConverter.from_session函數將 keras 模型直接轉換為.tflite 。 將下面的代碼fit_generator之后導出（用tensorflow 1.3.1測試）

with tf.keras.backend.get_session() as sess:
    sess.run(tf.global_variables_initializer())    
    converter = tf.lite.TFLiteConverter.from_session(sess, model.inputs, model.outputs)
    tflite_model = converter.convert()
    with open("model.tflite", "wb") as f:
        f.write(tflite_model)   

Answer 2

參加聚會有點晚了，但您可以這樣做：

converter = tf.lite.TFLiteConverter.from_keras_model(model)
tflite_model = converter.convert()

來源： https : //www.tensorflow.org/lite/convert/python_api