6-1构建模型的3种方法
可以使用以下3种方式构建模型：
- 使用Sequential按层顺序构建模型;
- 使用函数式API构建任意结构模型;
- 继承Model基类构建自定义模型。
对于顺序结构的模型，优先使用Sequential方法构建。
如果模型有多输入或者多输出，或者模型需要共享权重，或者模型具有残差连接等非顺序结构，推荐使用函数式API进行创建。
如果无特定必要，尽可能避免使用Model子类化的方式构建模型，这种方式提供了极大的灵活性，但也有更大的概率出错。
下面以IMDB电影评论的分类问题为例，演示3种创建模型的方法。
import numpy as np
import pandas as pd
import tensorflow as tf
from tqdm import tqdm
from tensorflow.keras import *
train_token_path = "../一、TensorFlow的建模流程/data/imdb/train_token.csv"
test_token_path = "../一、TensorFlow的建模流程/data/imdb/test_token.csv"
MAX_WORDS = 10000 # We will only consider the top 10000 words in the dataset
MAX_LEN = 200 # We will cut reviews after 200 words
BATCH_SIZE = 20
# 构建管道
def parse_line(line):
t = tf.strings.split(line, "\t")
label = tf.reshape(tf.cast(tf.strings.to_number(t[0]), tf.int32), (-1,))
features = tf.cast(tf.strings.to_number(tf.strings.split(t[1], " ")), tf.int32)
return features, label
ds_train = tf.data.TextLineDataset(filenames=[train_token_path]) \
.map(parse_line, num_parallel_calls=tf.data.experimental.AUTOTUNE) \
.shuffle(buffer_size=1000).batch(BATCH_SIZE) \
.prefetch(tf.data.experimental.AUTOTUNE)
ds_test = tf.data.TextLineDataset(filenames=[test_token_path]) \
.map(parse_line, num_parallel_calls=tf.data.experimental.AUTOTUNE) \
.shuffle(buffer_size=1000).batch(BATCH_SIZE) \
.prefetch(tf.data.experimental.AUTOTUNE)
Sequential按层顺序创建模型
tf.keras.backend.clear_session()
model = models.Sequential()
model.add(layers.Embedding(MAX_WORDS, 7, input_length=MAX_LEN))
model.add(layers.Conv1D(filters=64, kernel_size=5, activation="relu")) # kernel_size: An integer or tuple/list of a single integer, specifying the length of the 1D convolution window.
model.add(layers.MaxPool1D(2))
model.add(layers.Conv1D(filters=32, kernel_size=3, activation="relu"))
model.add(layers.MaxPool1D(2))
model.add(layers.Flatten())
model.add(layers.Dense(1, activation="sigmoid"))
model.compile(optimizer='Nadam', loss="binary_crossentropy", metrics=['accuracy', 'AUC'])
model.summary()
"""
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding (Embedding) (None, 200, 7) 70000
_________________________________________________________________
conv1d (Conv1D) (None, 196, 64) 2304
_________________________________________________________________
max_pooling1d (MaxPooling1D) (None, 98, 64) 0
_________________________________________________________________
conv1d_1 (Conv1D) (None, 96, 32) 6176
_________________________________________________________________
max_pooling1d_1 (MaxPooling1 (None, 48, 32) 0
_________________________________________________________________
flatten (Flatten) (None, 1536) 0
_________________________________________________________________
dense (Dense) (None, 1) 1537
=================================================================
Total params: 80,017
Trainable params: 80,017
Non-trainable params: 0
_________________________________________________________________
"""
import datetime
baselogger = callbacks.BaseLogger(stateful_metrics=['AUC'])
logdir = "./data/keras_model/" + datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
tensorboard_callback = tf.keras.callbacks.TensorBoard(logdir, histogram_freq=1)
history = model.fit(ds_train, validation_data=ds_test, epochs=6, callbacks=[tensorboard_callback]) # [baselogger, tensorboard_callback]
"""
Epoch 1/6
1000/1000 [==============================] - 5s 4ms/step - loss: 0.4746 - accuracy: 0.7350 - auc: 0.8418 - val_loss: 0.3161 - val_accuracy: 0.8632 - val_auc: 0.9411
Epoch 2/6
1000/1000 [==============================] - 4s 4ms/step - loss: 0.2400 - accuracy: 0.9049 - auc: 0.9650 - val_loss: 0.3084 - val_accuracy: 0.8768 - val_auc: 0.9466
Epoch 3/6
1000/1000 [==============================] - 4s 4ms/step - loss: 0.1628 - accuracy: 0.9392 - auc: 0.9835 - val_loss: 0.3503 - val_accuracy: 0.8702 - val_auc: 0.9412
Epoch 4/6
1000/1000 [==============================] - 5s 5ms/step - loss: 0.1069 - accuracy: 0.9618 - auc: 0.9924 - val_loss: 0.4766 - val_accuracy: 0.8596 - val_auc: 0.9296
Epoch 5/6
1000/1000 [==============================] - 8s 8ms/step - loss: 0.0645 - accuracy: 0.9779 - auc: 0.9971 - val_loss: 0.6010 - val_accuracy: 0.8600 - val_auc: 0.9221
Epoch 6/6
1000/1000 [==============================] - 7s 7ms/step - loss: 0.0354 - accuracy: 0.9877 - auc: 0.9990 - val_loss: 0.7835 - val_accuracy: 0.8556 - val_auc: 0.9111
"""
%matplotlib inline
%config InlineBackend.figure_format = 'svg'
import matplotlib.pyplot as plt
def plot_metric(history, metric):
train_metrics = history.history[metric]
val_metrics = history.history['val_' + metric]
epochs = range(1, len(train_metrics) + 1)
plt.plot(epochs, train_metrics, 'bo--')
plt.plot(epochs, val_metrics, 'ro-')
plt.title('Training and validation' + metric)
plt.xlabel("Epochs")
plt.ylabel(metric)
plt.legend(['train_' + metric, 'val_' + metric])
plt.show()
plot_metric(history, "auc")
函数式API创建任意结构模型
tf.keras.backend.clear_session()
inputs = layers.Input(shape=[MAX_LEN])
x = layers.Embedding(MAX_WORDS, 7)(inputs)
branch1 = layers.SeparableConv1D(64, 3, activation="relu")(x) # 可分离卷积，用不同的卷积核在各个通道上进行卷积，然后再用一个卷积核将不同的通道合并
branch1 = layers.MaxPool1D(3)(branch1)
branch1 = layers.SeparableConv1D(32, 3, activation="relu")(branch1)
branch1 = layers.GlobalMaxPool1D()(branch1)
branch2 = layers.SeparableConv1D(64, 5, activation="relu")(x) # 可分离卷积，用不同的卷积核在各个通道上进行卷积，然后再用一个卷积核将不同的通道合并
branch2 = layers.MaxPool1D(3)(branch2)
branch2 = layers.SeparableConv1D(32, 5, activation="relu")(branch2)
branch2 = layers.GlobalMaxPool1D()(branch2)
branch3 = layers.SeparableConv1D(64, 7, activation="relu")(x) # 可分离卷积，用不同的卷积核在各个通道上进行卷积，然后再用一个卷积核将不同的通道合并
branch3 = layers.MaxPool1D(3)(branch3)
branch3 = layers.SeparableConv1D(32, 7, activation="relu")(branch3)
branch3 = layers.GlobalMaxPool1D()(branch3)
concat = layers.Concatenate()([branch1, branch2, branch3])
outputs = layers.Dense(1, activation="sigmoid")(concat)
model = models.Model(inputs=inputs, outputs=outputs)
model.compile(optimizer="Nadam", loss="binary_crossentropy", metrics=["accuracy", "AUC"])
model.summary()
"""
Model: "model"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_1 (InputLayer) [(None, 200)] 0
__________________________________________________________________________________________________
embedding (Embedding) (None, 200, 7) 70000 input_1[0][0]
__________________________________________________________________________________________________
separable_conv1d (SeparableConv (None, 198, 64) 533 embedding[0][0]
__________________________________________________________________________________________________
separable_conv1d_2 (SeparableCo (None, 196, 64) 547 embedding[0][0]
__________________________________________________________________________________________________
separable_conv1d_4 (SeparableCo (None, 194, 64) 561 embedding[0][0]
__________________________________________________________________________________________________
max_pooling1d (MaxPooling1D) (None, 66, 64) 0 separable_conv1d[0][0]
__________________________________________________________________________________________________
max_pooling1d_1 (MaxPooling1D) (None, 65, 64) 0 separable_conv1d_2[0][0]
__________________________________________________________________________________________________
max_pooling1d_2 (MaxPooling1D) (None, 64, 64) 0 separable_conv1d_4[0][0]
__________________________________________________________________________________________________
separable_conv1d_1 (SeparableCo (None, 64, 32) 2272 max_pooling1d[0][0]
__________________________________________________________________________________________________
separable_conv1d_3 (SeparableCo (None, 61, 32) 2400 max_pooling1d_1[0][0]
__________________________________________________________________________________________________
separable_conv1d_5 (SeparableCo (None, 58, 32) 2528 max_pooling1d_2[0][0]
__________________________________________________________________________________________________
global_max_pooling1d (GlobalMax (None, 32) 0 separable_conv1d_1[0][0]
__________________________________________________________________________________________________
global_max_pooling1d_1 (GlobalM (None, 32) 0 separable_conv1d_3[0][0]
__________________________________________________________________________________________________
global_max_pooling1d_2 (GlobalM (None, 32) 0 separable_conv1d_5[0][0]
__________________________________________________________________________________________________
concatenate (Concatenate) (None, 96) 0 global_max_pooling1d[0][0]
global_max_pooling1d_1[0][0]
global_max_pooling1d_2[0][0]
__________________________________________________________________________________________________
dense (Dense) (None, 1) 97 concatenate[0][0]
==================================================================================================
Total params: 78,938
Trainable params: 78,938
Non-trainable params: 0
__________________________________________________________________________________________________
"""
import datetime
logdir = "./data/keras_model/" + datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
tensorboard_callback = tf.keras.callbacks.TensorBoard(logdir, histogram_freq=1)
history = model.fit(ds_train, validation_data=ds_test, epochs=6, callbacks=[tensorboard_callback])
"""
Epoch 1/6
1000/1000 [==============================] - 18s 17ms/step - loss: 0.5480 - accuracy: 0.6829 - auc: 0.7781 - val_loss: 0.3778 - val_accuracy: 0.8328 - val_auc: 0.9134
Epoch 2/6
1000/1000 [==============================] - 17s 17ms/step - loss: 0.3190 - accuracy: 0.8637 - auc: 0.9379 - val_loss: 0.3464 - val_accuracy: 0.8462 - val_auc: 0.9275
Epoch 3/6
1000/1000 [==============================] - 17s 17ms/step - loss: 0.2331 - accuracy: 0.9089 - auc: 0.9669 - val_loss: 0.3582 - val_accuracy: 0.8510 - val_auc: 0.9288
Epoch 4/6
1000/1000 [==============================] - 14s 14ms/step - loss: 0.1680 - accuracy: 0.9391 - auc: 0.9824 - val_loss: 0.4038 - val_accuracy: 0.8484 - val_auc: 0.9271
Epoch 5/6
1000/1000 [==============================] - 8s 8ms/step - loss: 0.1134 - accuracy: 0.9617 - auc: 0.9915 - val_loss: 0.4620 - val_accuracy: 0.8484 - val_auc: 0.9243
Epoch 6/6
1000/1000 [==============================] - 12s 12ms/step - loss: 0.0678 - accuracy: 0.9793 - auc: 0.9966 - val_loss: 0.5725 - val_accuracy: 0.8434 - val_auc: 0.9163
"""
plot_metric(history, "auc")
Model子类化创建自定义模型
# 先自定义一个残差模块，为自定义Layer
class ResBlock(layers.Layer):
def __init__(self, kernel_size, **kwargs):
super().__init__(**kwargs)
self.kernel_size = kernel_size
def build(self, input_shape):
self.conv1 = layers.Conv1D(filters=64, kernel_size=self.kernel_size, activation="relu", padding="same")
self.conv2 = layers.Conv1D(filters=32, kernel_size=self.kernel_size, activation="relu", padding="same")
self.conv3 = layers.Conv1D(filters=input_shape[-1], kernel_size=self.kernel_size, activation="relu", padding="same")
self.maxpool = layers.MaxPool1D(2)
super().build(input_shape) # 相当于设置self.built=True
def call(self, inputs):
x = self.conv1(inputs)
x = self.conv2(x)
x = self.conv3(x)
x = layers.Add()([inputs, x])
x = self.maxpool(x)
return x
# 如果要让自定义的Layer通过Functional API组合成模型时可以序列化，需要自定义 get_config方法
def get_config(self):
config = super().get_config()
config.update({"kernel_size": self.kernel_size})
return config
# 测试ResBlock
resblock = ResBlock(kernel_size=3)
resblock.build(input_shape=(None, 200, 7))
resblock.compute_output_shape(input_shape=(None, 200, 7))
"""
TensorShape([None, 100, 7])
"""
# 自定义模型，实际上也可以使用Sequential或者Functional API
class ImdbModel(models.Model):
def __init__(self):
super().__init__()
def build(self, input_shape):
self.embedding = layers.Embedding(MAX_WORDS, 7)
self.block1 = ResBlock(7)
self.block2 = ResBlock(5)
self.dense = layers.Dense(1, activation="sigmoid")
super().build(input_shape)
def call(self, x):
x = self.embedding(x)
x = self.block1(x)
x = self.block2(x)
x = layers.Flatten()(x)
x = self.dense(x)
return x
tf.keras.backend.clear_session()
model = ImdbModel()
model.build(input_shape=(None, 200))
model.summary()
model.compile(optimizer="Nadam", loss="binary_crossentropy", metrics=["accuracy", "AUC"])
"""
Model: "imdb_model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding (Embedding) multiple 70000
_________________________________________________________________
res_block (ResBlock) multiple 19143
_________________________________________________________________
res_block_1 (ResBlock) multiple 13703
_________________________________________________________________
dense (Dense) multiple 351
=================================================================
Total params: 103,197
Trainable params: 103,197
Non-trainable params: 0
_________________________________________________________________
"""
import datetime
logdir = "./data/tflogs/keras_model/" + datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
tensorboard_callback = tf.keras.callbacks.TensorBoard(logdir, histogram_freq=1)
history = model.fit(ds_train, validation_data=ds_test, epochs=6, callbacks=[tensorboard_callback])
"""
Epoch 1/6
1000/1000 [==============================] - 18s 17ms/step - loss: 0.4992 - accuracy: 0.7184 - auc: 0.8215 - val_loss: 0.3192 - val_accuracy: 0.8588 - val_auc: 0.9379
Epoch 2/6
1000/1000 [==============================] - 18s 18ms/step - loss: 0.2512 - accuracy: 0.8989 - auc: 0.9618 - val_loss: 0.3416 - val_accuracy: 0.8594 - val_auc: 0.9449
Epoch 3/6
1000/1000 [==============================] - 17s 17ms/step - loss: 0.1651 - accuracy: 0.9369 - auc: 0.9831 - val_loss: 0.3654 - val_accuracy: 0.8710 - val_auc: 0.9412
Epoch 4/6
1000/1000 [==============================] - 17s 17ms/step - loss: 0.1037 - accuracy: 0.9611 - auc: 0.9930 - val_loss: 0.4717 - val_accuracy: 0.8676 - val_auc: 0.9338
Epoch 5/6
1000/1000 [==============================] - 17s 17ms/step - loss: 0.0606 - accuracy: 0.9777 - auc: 0.9973 - val_loss: 0.5980 - val_accuracy: 0.8584 - val_auc: 0.9230
Epoch 6/6
1000/1000 [==============================] - 17s 17ms/step - loss: 0.0320 - accuracy: 0.9894 - auc: 0.9991 - val_loss: 0.9280 - val_accuracy: 0.8456 - val_auc: 0.9062
"""
plot_metric(history, "auc")
