keras overviewer

08 Jun 2020

1. BUild Modle

1.1 Sequential model

模型通常是graphy of layers，最通用的是sequential model，即堆叠层模型

from tensorflow.keras import layers

model = tf.keras.Sequential()
# Adds a densely-connected layer with 64 units to the model:
# 64为隐藏单元个数，relu是激活函数
model.add(layers.Dense(64, activation='relu'))
# Add another:
model.add(layers.Dense(64, activation='relu'))
# Add an output layer with 10 output units:
model.add(layers.Dense(10))

1.2 configure layers

一些通用的构造参数：

• activation：激活函数，可以设置为内建函数或者callable abject；默认不设置激活函数
• kernel_initializer and bias_initializer：初始化层的权值（kernel和bias）。 This parameter is a name or a callable object；默认是the "Glorot uniform" initializer, and the bias defaults to zeros.
• kernel_regularizer and bias_regularizer：对权值正则化，如L1或L2；默认没有

# Create a relu layer:
layers.Dense(64, activation='relu')
# Or:
layers.Dense(64, activation=tf.nn.relu)

# A linear layer with L1 regularization of factor 0.01 applied to the kernel matrix:
layers.Dense(64, kernel_regularizer=tf.keras.regularizers.l1(0.01))

# A linear layer with L2 regularization of factor 0.01 applied to the bias vector:
layers.Dense(64, bias_regularizer=tf.keras.regularizers.l2(0.01))

# A linear layer with a kernel initialized to a random orthogonal matrix:
layers.Dense(64, kernel_initializer='orthogonal')

# A linear layer with a bias vector initialized to 2.0s:
layers.Dense(64, bias_initializer=tf.keras.initializers.Constant(2.0))

2. 训练和估计

2.1 训练

构建好模型后，用compile方法调整学习过程；

model = tf.keras.Sequential([
# Adds a densely-connected layer with 64 units to the model:
layers.Dense(64, activation='relu', input_shape=(32,)),
# Add another:
layers.Dense(64, activation='relu'),
# Add an output layer with 10 output units:
layers.Dense(10)])

model.compile(optimizer=tf.keras.optimizers.Adam(0.01),
              loss=tf.keras.losses.CategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])

compile有三个重要参数：

• optimizer：Pass it optimizer instances from the tf.keras.optimizers module, such as tf.keras.optimizers.Adam or tf.keras.optimizers.SGD；
• loss：优化过程中的最小化函数。通用选项包含均方误差(mse)；categorical_crossentropy, and binary_crossentropy. Loss functions are specified by name or by passing a callable object from the tf.keras.losses module.
• metrics: 监视训练过程. These are string names or callables from the tf.keras.metrics module
• Additionally, to make sure the model trains and evaluates eagerly, you can make sure to pass run_eagerly=True as a parameter to compile.

3. 预测

4. 其他复杂模型

4.2 其他模型

除了sequential模型，还有多输出多输出模型，共享层模型，非堆叠模型（residual connections）；

参考：Keras–overviewer