tensorflow2教程17：文本卷积
1.基本文本卷积
For more information, refer to:
Kim 2014 (http://emnlp2014.org/papers/pdf/EMNLP2014181.pdf)
Zhang et al 2015 (https://papers.nips.cc/paper/5782-character-level-convolutional-networks-for-text-classification.pdf)
- 使用卷积进行句子分类(Kim 2014)
- 多个卷积核
import numpy as np
import matplotlib.pyplot as plt
from tensorflow import keras
from tensorflow.keras import layers
from tensorflow.keras.preprocessing.sequence import pad_sequences
num_features = 3000
sequence_length = 300
embedding_dimension = 100
(x_train, y_train), (x_test, y_test) = keras.datasets.imdb.load_data(num_words=num_features)
print(x_train.shape)
print(x_test.shape)
print(y_train.shape)
print(y_test.shape)
(25000,)
(25000,)
(25000,)
(25000,)
x_train = pad_sequences(x_train, maxlen=sequence_length)
x_test = pad_sequences(x_test, maxlen=sequence_length)
print(x_train.shape)
print(x_test.shape)
print(y_train.shape)
print(y_test.shape)
(25000, 300)
(25000, 300)
(25000,)
(25000,)
2.构造基本句子分类器
def imdb_cnn():
model = keras.Sequential([
layers.Embedding(input_dim=num_features, output_dim=embedding_dimension,input_length=sequence_length),
layers.Conv1D(filters=50, kernel_size=5, strides=1, padding='valid'),
layers.MaxPool1D(2, padding='valid'),
layers.Flatten(),
layers.Dense(10, activation='relu'),
layers.Dense(1, activation='sigmoid')
])
model.compile(optimizer=keras.optimizers.Adam(1e-3),
loss=keras.losses.BinaryCrossentropy(),
metrics=['accuracy'])
return model
model = imdb_cnn()
model.summary()
Model: "sequential_3"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding_3 (Embedding) (None, 300, 100) 300000
_________________________________________________________________
conv1d_3 (Conv1D) (None, 296, 50) 25050
_________________________________________________________________
max_pooling1d_3 (MaxPooling1 (None, 148, 50) 0
_________________________________________________________________
flatten_3 (Flatten) (None, 7400) 0
_________________________________________________________________
dense_6 (Dense) (None, 10) 74010
_________________________________________________________________
dense_7 (Dense) (None, 1) 11
=================================================================
Total params: 399,071
Trainable params: 399,071
Non-trainable params: 0
_________________________________________________________________
%%time
history = model.fit(x_train, y_train, batch_size=64, epochs=5, validation_split=0.1)
Train on 22500 samples, validate on 2500 samples
Epoch 1/5
22500/22500 [==============================] - 23s 1ms/sample - loss: 0.4255 - accuracy: 0.7789 - val_loss: 0.3002 - val_accuracy: 0.8732
Epoch 2/5
22500/22500 [==============================] - 22s 960us/sample - loss: 0.2367 - accuracy: 0.9060 - val_loss: 0.2985 - val_accuracy: 0.8752
Epoch 3/5
22500/22500 [==============================] - 23s 1ms/sample - loss: 0.1171 - accuracy: 0.9603 - val_loss: 0.3549 - val_accuracy: 0.8680
Epoch 4/5
22500/22500 [==============================] - 22s 969us/sample - loss: 0.0286 - accuracy: 0.9937 - val_loss: 0.4745 - val_accuracy: 0.8656
Epoch 5/5
22500/22500 [==============================] - 22s 964us/sample - loss: 0.0047 - accuracy: 0.9996 - val_loss: 0.5525 - val_accuracy: 0.8684
CPU times: user 4min 33s, sys: 4.77 s, total: 4min 38s
Wall time: 1min 50s
plt.plot(history.history['accuracy'])
plt.plot(history.history['val_accuracy'])
plt.legend(['training', 'valiation'], loc='upper left')
plt.show()
3.多核卷积网络
filter_sizes=[3,4,5]
def convolution():
inn = layers.Input(shape=(sequence_length, embedding_dimension, 1))
cnns = []
for size in filter_sizes:
conv = layers.Conv2D(filters=64, kernel_size=(size, embedding_dimension),
strides=1, padding='valid', activation='relu')(inn)
pool = layers.MaxPool2D(pool_size=(sequence_length-size+1, 1), padding='valid')(conv)
cnns.append(pool)
outt = layers.concatenate(cnns)
model = keras.Model(inputs=inn, outputs=outt)
return model
def cnn_mulfilter():
model = keras.Sequential([
layers.Embedding(input_dim=num_features, output_dim=embedding_dimension,
input_length=sequence_length),
layers.Reshape((sequence_length, embedding_dimension, 1)),
convolution(),
layers.Flatten(),
layers.Dense(10, activation='relu'),
layers.Dropout(0.2),
layers.Dense(1, activation='sigmoid')
])无锡妇科医院排行 http://www.0510bhyy.com/
model.compile(optimizer=keras.optimizers.Adam(),
loss=keras.losses.BinaryCrossentropy(),
metrics=['accuracy'])
return model
model = cnn_mulfilter()
model.summary()
Model: "sequential_4"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding_5 (Embedding) (None, 300, 100) 300000
_________________________________________________________________
reshape_1 (Reshape) (None, 300, 100, 1) 0
_________________________________________________________________
model (Model) (None, 1, 1, 192) 76992
_________________________________________________________________
flatten_4 (Flatten) (None, 192) 0
_________________________________________________________________
dense_8 (Dense) (None, 10) 1930
_________________________________________________________________
dropout (Dropout) (None, 10) 0
_________________________________________________________________
dense_9 (Dense) (None, 1) 11
=================================================================
Total params: 378,933
Trainable params: 378,933
Non-trainable params: 0
_________________________________________________________________
%%time
history = model.fit(x_train, y_train, batch_size=64, epochs=5, validation_split=0.1)
Train on 22500 samples, validate on 2500 samples
Epoch 1/5
22500/22500 [==============================] - 40s 2ms/sample - loss: 0.4740 - accuracy: 0.7557 - val_loss: 0.3136 - val_accuracy: 0.8724
Epoch 2/5
22500/22500 [==============================] - 40s 2ms/sample - loss: 0.2623 - accuracy: 0.8968 - val_loss: 0.2967 - val_accuracy: 0.8796
Epoch 3/5
22500/22500 [==============================] - 43s 2ms/sample - loss: 0.1840 - accuracy: 0.9350 - val_loss: 0.2893 - val_accuracy: 0.8868
Epoch 4/5
22500/22500 [==============================] - 40s 2ms/sample - loss: 0.1215 - accuracy: 0.9599 - val_loss: 0.3308 - val_accuracy: 0.8772
Epoch 5/5
22500/22500 [==============================] - 45s 2ms/sample - loss: 0.0694 - accuracy: 0.9806 - val_loss: 0.3699 - val_accuracy: 0.8848
CPU times: user 10min 52s, sys: 4.88 s, total: 10min 56s
Wall time: 3min 28s
plt.plot(history.history['accuracy'])
plt.plot(history.history['val_accuracy'])
plt.legend(['training', 'valiation'], loc='upper left')
plt.show()
