TensorFlow model

两种创建model的方式
1:链式函数创建
要创建输入层inputs

import tensorflow as tf

inputs = tf.keras.Input(shape=(3,))
x = tf.keras.layers.Dense(4, activation=tf.nn.relu)(inputs)
outputs = tf.keras.layers.Dense(5, activation=tf.nn.softmax)(x)
model = tf.keras.Model(inputs=inputs, outputs=outputs)

2:使用对象创建

import tensorflow as tf

class MyModel(tf.keras.Model):

  def __init__(self):
    super(MyModel, self).__init__()
    self.dense1 = tf.keras.layers.Dense(4, activation=tf.nn.relu)
    self.dense2 = tf.keras.layers.Dense(5, activation=tf.nn.softmax)

  def call(self, inputs):
    x = self.dense1(inputs)
    return self.dense2(x)

model = MyModel()

属性

属性	描述
layers	层
metrics_names	所有输出的标签
run_eagerly	是否使用eagerly模式,默认False,静态图
sample_weights
state_updates

方法

1. compile

compile(
    optimizer,
    loss=None,
    metrics=None,
    loss_weights=None,
    sample_weight_mode=None,
    weighted_metrics=None,
    target_tensors=None,
    distribute=None,
    **kwargs
)

参数	描述
optimizer	(string,Object)优化器
loss	(String,Object,Function),如果模型有多个输出,可以为不同的输出指定不同的损失函数
metrics	(List(String))衡量指标,比如[‘accuracy’,‘mse’]
loss_weights
sample_weight_mode
weighted_metrics
target_tensors
distribute
**kwargs

evaluate

evaluate(
    x=None,
    y=None,
    batch_size=None,
    verbose=1,
    sample_weight=None,
    steps=None,
    callbacks=None,
    max_queue_size=10,
    workers=1,
    use_multiprocessing=False
)

参数	描述
x	(numpy array;tensor;[tensor];dict;tf.data;keras.utils.Sequence)
y
batch_size	(int)每一次梯度下降使用的样本数量.默认为32,如果输入数据已经指定了batch_size,则不要再次指定
verbose
sample_weight
steps	(int)执行多少个batch之后打印日志信息,默认,一个epoch,打印一次
callbacks
max_queue_size
workers
use_multiprocessing

evaluate_generator

evaluate_generator(
    generator,
    steps=None,
    callbacks=None,
    max_queue_size=10,
    workers=1,
    use_multiprocessing=False,
    verbose=0
)

fit

fit_generator

get_layer

get_layer(
    name=None,
    index=None
)

load_weights

load_weights(
    filepath,
    by_name=False
)

predict

predict(
    x,
    batch_size=None,
    verbose=0,
    steps=None,
    callbacks=None,
    max_queue_size=10,
    workers=1,
    use_multiprocessing=False
)

predict_generator

predict_generator(
    generator,
    steps=None,
    callbacks=None,
    max_queue_size=10,
    workers=1,
    use_multiprocessing=False,
    verbose=0
)

predict_on_batch

predict_on_batch(x)

reset_metrics

reset_states

save

保存模型为HDF5文件

save_weights

summary

summary(
    line_length=None,
    positions=None,
    print_fn=None
)

test_on_batch

to_json

to_yaml

train_on_batch

参考：
https://tensorflow.google.cn/versions/r2.0/api_docs/python/tf/keras/Model