mnist_tensorboard_embedding_visualization.py

from os.path import join
import keras
from keras.datasets import mnist
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras import backend as K
import numpy as np

batch_size = 128
num_classes = 10
epochs = 12

#建立資料夾，可以將callback儲存的log丟進來
log_dir = 'C:\\Users\\Lido_Lee\\Downloads\\google_mnist_ex'

# input image dimensions
img_rows, img_cols = 28, 28

# 直接從 Keras data 庫讀取 MNIST data
(x_train, y_train), (x_test, y_test) = mnist.load_data()

#判斷 Keras 後端讀取資料格式
if K.image_data_format() == 'channels_first':
    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)
    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)
    input_shape = (1, img_rows, img_cols)
else:
    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)
    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)
    input_shape = (img_rows, img_cols, 1)

#數據預處理，轉格float32格式，且值在0~1之間
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255
print('x_train shape:', x_train.shape)
print(x_train.shape[0], 'train samples')
print(x_test.shape[0], 'test samples')

# 將類標籤保存到metadata
with open(join(log_dir, 'metadata.tsv'), 'w') as f:
    np.savetxt(f, y_test)

# 標籤預處理，用Keras to_categorical 做分類標籤(10個分類)
y_train = keras.utils.to_categorical(y_train, num_classes)
y_test = keras.utils.to_categorical(y_test, num_classes)

#建立模型
model = Sequential()
model.add(Conv2D(32, kernel_size=(3, 3),activation='relu',input_shape=input_shape))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(128, activation='relu', name='features'))
model.add(Dropout(0.5))
model.add(Dense(num_classes, activation='softmax'))

#建立編譯器，選擇優化器、損失函數、指標
model.compile(optimizer=keras.optimizers.Adadelta(),
              loss=keras.losses.categorical_crossentropy,
              metrics=['accuracy'])

#回調函數，這邊使用EarlyStopping、TensorBoard
callbacks = [
    keras.callbacks.EarlyStopping(monitor='acc',
                                  patience=1, ),
    keras.callbacks.TensorBoard(log_dir=log_dir,
                                histogram_freq=1,
                                batch_size=batch_size,
                                embeddings_freq=1,
                                write_graph=False,
                                embeddings_layer_names=['features'],
                                embeddings_metadata=['metadata.tsv'],
                                embeddings_data=x_test)
            ]
#丟進模型訓練，並調用 callbacks 儲存 log
history = model.fit(x_train, y_train,
                    batch_size=batch_size,
                    callbacks=callbacks,
                    epochs=epochs,
                    verbose=1,
                    validation_split=0.2
                    )

#驗證測試資料
score = model.evaluate(x_test, y_test, verbose=0)
print('Test loss:', score[0])
print('Test accuracy:', score[1])