基于tensorflow_gpu 1.9.0实现的第二个神经网络： 对影评的二值分类
示例来源于Tensorflow的官方教程。
基于tensorflow_gpu 1.9.0实现的第二个神经网络：对影评的二值分类，代码如下：
#!/usr/bin/env python
import tensorflow as tf
from tensorflow import keras
import numpy as np
import matplotlib.pyplot as plt
print(tf.__version__)
# Download the IMDB dataset
imdb = keras.datasets.imdb
(train_data, train_labels), (test_data,
test_labels) = imdb.load_data(num_words=10000)
# Explore the data
print("Training entries: {}, labels: {}".format(
len(train_data), len(train_labels)))
print(train_data[0])
len(train_data[0]), len(train_data[1])
# A dictionary mapping words to an integer index
word_index = imdb.get_word_index()
# The first indices are reserved
word_index = {k: (v+3) for k, v in word_index.items()}
word_index["" ] = 0
word_index["" ] = 1
word_index["" ] = 2 # unknown
word_index["UNUSED"] = 3
reverse_word_index = dict([(value, key)
for (key, value) in word_index.items()])
def decode_review(text):
return ' '.join([reverse_word_index.get(i, '?') for i in text])
decode_review(train_data[0])
# Prepare the data
# The reviews--the arrays of integers--must be converted to tensors before fed
# into the neural network.
# Since the movie reviews must be the same length, we will use the pad_sequences
# function to standardize the lengths.
train_data = keras.preprocessing.sequence.pad_sequences(train_data,
value=word_index["" ],
padding='post',
maxlen=256)
test_data = keras.preprocessing.sequence.pad_sequences(test_data,
value=word_index["" ],
padding='post',
maxlen=256)
len(train_data[0]), len(train_data[1])
print(train_data[0])
# Build the model
# Input shape is the vocabulary count used for the movie reviews (10, 000 words)
vocab_size = 10000
model = keras.Sequential()
model.add(keras.layers.Embedding(vocab_size, 16))
model.add(keras.layers.GlobalAveragePooling1D())
model.add(keras.layers.Dense(16, activation=tf.nn.relu))
model.add(keras.layers.Dense(1, activation=tf.nn.sigmoid))
model.summary()
# Loss function and optimizer
model.compile(optimizer=tf.train.AdamOptimizer(),
loss='binary_crossentropy',
metrics=['accuracy'])
# Create a validation set
x_val = train_data[:10000]
partial_x_train = train_data[10000:]
y_val = train_labels[:10000]
partial_y_train = train_labels[10000:]
# Train the model
history = model.fit(partial_x_train,
partial_y_train,
epochs=40,
batch_size=512,
validation_data=(x_val, y_val),
verbose=1)
# Evaluate the model
results = model.evaluate(test_data, test_labels)
print(results)
# Create a graph of accuracy and loss over time
history_dict = history.history
history_dict.keys()
acc = history_dict['acc']
val_acc = history_dict['val_acc']
loss = history_dict['loss']
val_loss = history_dict['val_loss']
epochs = range(1, len(acc) + 1)
plt.figure()
# "bo" is for "blue dot"
plt.plot(epochs, loss, 'bo', label='Training loss')
# "b" is for "solid blue line"
plt.plot(epochs, val_loss, 'b', label='Validattion loss')
plt.title('Training and validation loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.legend()
plt.figure()
# "bo" is for "blue dot"
plt.plot(epochs, acc, 'bo', label='Training accuracy')
# "b" is for "solid blue line"
plt.plot(epochs, val_acc, 'b', label='Validattion accuracy')
plt.title('Training and validation accuracy')
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.legend()
plt.show()
运行该程序之前，需要先根据我的另一篇博客《Ubuntu 16.04安装tensorflow_gpu 1.9.0的方法》，安装tensorflow_gpu 1.9.0，然后安装numpy、matplotlib等工具包，此外还需要安装python-tk，命令如下：
sudo pip install numpy
sudo pip install matplotlib
sudo apt-get install python-tk
运行结果如下：
1.9.0
Training entries: 25000, labels: 25000
[1, 14, 22, 16, 43, 530, 973, 1622, 1385, 65, 458, 4468, 66, 3941, 4, 173, 36, 256, 5, 25, 100, 43, 838, 112, 50, 670, 2, 9, 35, 480, 284, 5, 150, 4, 172, 112, 167, 2, 336, 385, 39, 4, 172, 4536, 1111, 17, 546, 38, 13, 447, 4, 192, 50, 16, 6, 147, 2025, 19, 14, 22, 4, 1920, 4613, 469, 4, 22, 71, 87, 12, 16, 43, 530, 38, 76, 15, 13, 1247, 4, 22, 17, 515, 17, 12, 16, 626, 18, 2, 5, 62, 386, 12, 8, 316, 8, 106, 5, 4, 2223, 5244, 16, 480, 66, 3785, 33, 4, 130, 12, 16, 38, 619, 5, 25, 124, 51, 36, 135, 48, 25, 1415, 33, 6, 22, 12, 215, 28, 77, 52, 5, 14, 407, 16, 82, 2, 8, 4, 107, 117, 5952, 15, 256, 4, 2, 7, 3766, 5, 723, 36, 71, 43, 530, 476, 26, 400, 317, 46, 7, 4, 2, 1029, 13, 104, 88, 4, 381, 15, 297, 98, 32, 2071, 56, 26, 141, 6, 194, 7486, 18, 4, 226, 22, 21, 134, 476, 26, 480, 5, 144, 30, 5535, 18, 51, 36, 28, 224, 92, 25, 104, 4, 226, 65, 16, 38, 1334, 88, 12, 16, 283, 5, 16, 4472, 113, 103, 32, 15, 16, 5345, 19, 178, 32]
[ 1 14 22 16 43 530 973 1622 1385 65 458 4468 66 3941
4 173 36 256 5 25 100 43 838 112 50 670 2 9
35 480 284 5 150 4 172 112 167 2 336 385 39 4
172 4536 1111 17 546 38 13 447 4 192 50 16 6 147
2025 19 14 22 4 1920 4613 469 4 22 71 87 12 16
43 530 38 76 15 13 1247 4 22 17 515 17 12 16
626 18 2 5 62 386 12 8 316 8 106 5 4 2223
5244 16 480 66 3785 33 4 130 12 16 38 619 5 25
124 51 36 135 48 25 1415 33 6 22 12 215 28 77
52 5 14 407 16 82 2 8 4 107 117 5952 15 256
4 2 7 3766 5 723 36 71 43 530 476 26 400 317
46 7 4 2 1029 13 104 88 4 381 15 297 98 32
2071 56 26 141 6 194 7486 18 4 226 22 21 134 476
26 480 5 144 30 5535 18 51 36 28 224 92 25 104
4 226 65 16 38 1334 88 12 16 283 5 16 4472 113
103 32 15 16 5345 19 178 32 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0]
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding (Embedding) (None, None, 16) 160000
_________________________________________________________________
global_average_pooling1d (Gl (None, 16) 0
_________________________________________________________________
dense (Dense) (None, 16) 272
_________________________________________________________________
dense_1 (Dense) (None, 1) 17
=================================================================
Total params: 160,289
Trainable params: 160,289
Non-trainable params: 0
_________________________________________________________________
Train on 15000 samples, validate on 10000 samples
Epoch 1/40
2018-07-26 16:23:47.699236: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2018-07-26 16:23:47.754868: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:897] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2018-07-26 16:23:47.755179: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1392] Found device 0 with properties:
name: GeForce GTX 1050 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.392
pciBusID: 0000:01:00.0
totalMemory: 3.94GiB freeMemory: 2.92GiB
2018-07-26 16:23:47.755192: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1471] Adding visible gpu devices: 0
2018-07-26 16:23:47.927017: I tensorflow/core/common_runtime/gpu/gpu_device.cc:952] Device interconnect StreamExecutor with strength 1 edge matrix:
2018-07-26 16:23:47.927050: I tensorflow/core/common_runtime/gpu/gpu_device.cc:958] 0
2018-07-26 16:23:47.927060: I tensorflow/core/common_runtime/gpu/gpu_device.cc:971] 0: N
2018-07-26 16:23:47.927246: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1084] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 2635 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1050 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)
15000/15000 [==============================] - 1s 57us/step - loss: 0.6918 - acc: 0.5469 - val_loss: 0.6910 - val_acc: 0.5332
Epoch 2/40
15000/15000 [==============================] - 0s 20us/step - loss: 0.6896 - acc: 0.5797 - val_loss: 0.6885 - val_acc: 0.6327
Epoch 3/40
15000/15000 [==============================] - 0s 20us/step - loss: 0.6869 - acc: 0.6238 - val_loss: 0.6859 - val_acc: 0.6377
Epoch 4/40
15000/15000 [==============================] - 0s 18us/step - loss: 0.6829 - acc: 0.6443 - val_loss: 0.6821 - val_acc: 0.5562
Epoch 5/40
15000/15000 [==============================] - 0s 20us/step - loss: 0.6767 - acc: 0.7030 - val_loss: 0.6741 - val_acc: 0.7263
Epoch 6/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.6678 - acc: 0.7368 - val_loss: 0.6643 - val_acc: 0.7128
Epoch 7/40
15000/15000 [==============================] - 0s 18us/step - loss: 0.6551 - acc: 0.7505 - val_loss: 0.6508 - val_acc: 0.7452
Epoch 8/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.6381 - acc: 0.7613 - val_loss: 0.6337 - val_acc: 0.7558
Epoch 9/40
15000/15000 [==============================] - 0s 18us/step - loss: 0.6176 - acc: 0.7673 - val_loss: 0.6128 - val_acc: 0.7574
Epoch 10/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.5929 - acc: 0.7845 - val_loss: 0.5891 - val_acc: 0.7712
Epoch 11/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.5659 - acc: 0.7891 - val_loss: 0.5666 - val_acc: 0.7760
Epoch 12/40
15000/15000 [==============================] - 0s 18us/step - loss: 0.5382 - acc: 0.8040 - val_loss: 0.5381 - val_acc: 0.7920
Epoch 13/40
15000/15000 [==============================] - 0s 18us/step - loss: 0.5087 - acc: 0.8211 - val_loss: 0.5133 - val_acc: 0.8060
Epoch 14/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.4810 - acc: 0.8326 - val_loss: 0.4889 - val_acc: 0.8153
Epoch 15/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.4544 - acc: 0.8431 - val_loss: 0.4655 - val_acc: 0.8200
Epoch 16/40
15000/15000 [==============================] - 0s 18us/step - loss: 0.4299 - acc: 0.8486 - val_loss: 0.4442 - val_acc: 0.8329
Epoch 17/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.4061 - acc: 0.8603 - val_loss: 0.4252 - val_acc: 0.8394
Epoch 18/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.3851 - acc: 0.8675 - val_loss: 0.4080 - val_acc: 0.8444
Epoch 19/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.3664 - acc: 0.8734 - val_loss: 0.3931 - val_acc: 0.8474
Epoch 20/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.3494 - acc: 0.8794 - val_loss: 0.3799 - val_acc: 0.8534
Epoch 21/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.3341 - acc: 0.8845 - val_loss: 0.3684 - val_acc: 0.8582
Epoch 22/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.3203 - acc: 0.8880 - val_loss: 0.3584 - val_acc: 0.8624
Epoch 23/40
15000/15000 [==============================] - 0s 18us/step - loss: 0.3080 - acc: 0.8917 - val_loss: 0.3498 - val_acc: 0.8647
Epoch 24/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2963 - acc: 0.8957 - val_loss: 0.3418 - val_acc: 0.8682
Epoch 25/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2859 - acc: 0.9005 - val_loss: 0.3349 - val_acc: 0.8688
Epoch 26/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2762 - acc: 0.9036 - val_loss: 0.3292 - val_acc: 0.8728
Epoch 27/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2676 - acc: 0.9045 - val_loss: 0.3238 - val_acc: 0.8722
Epoch 28/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2588 - acc: 0.9087 - val_loss: 0.3185 - val_acc: 0.8740
Epoch 29/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2512 - acc: 0.9119 - val_loss: 0.3147 - val_acc: 0.8757
Epoch 30/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2447 - acc: 0.9129 - val_loss: 0.3106 - val_acc: 0.8755
Epoch 31/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2369 - acc: 0.9167 - val_loss: 0.3075 - val_acc: 0.8777
Epoch 32/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2310 - acc: 0.9175 - val_loss: 0.3045 - val_acc: 0.8779
Epoch 33/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2241 - acc: 0.9216 - val_loss: 0.3018 - val_acc: 0.8793
Epoch 34/40
15000/15000 [==============================] - 0s 20us/step - loss: 0.2184 - acc: 0.9229 - val_loss: 0.2999 - val_acc: 0.8786
Epoch 35/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.2133 - acc: 0.9237 - val_loss: 0.2973 - val_acc: 0.8820
Epoch 36/40
15000/15000 [==============================] - 0s 20us/step - loss: 0.2073 - acc: 0.9269 - val_loss: 0.2956 - val_acc: 0.8808
Epoch 37/40
15000/15000 [==============================] - 0s 20us/step - loss: 0.2022 - acc: 0.9286 - val_loss: 0.2940 - val_acc: 0.8804
Epoch 38/40
15000/15000 [==============================] - 0s 20us/step - loss: 0.1974 - acc: 0.9303 - val_loss: 0.2924 - val_acc: 0.8811
Epoch 39/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.1923 - acc: 0.9327 - val_loss: 0.2910 - val_acc: 0.8837
Epoch 40/40
15000/15000 [==============================] - 0s 19us/step - loss: 0.1878 - acc: 0.9338 - val_loss: 0.2898 - val_acc: 0.8836
25000/25000 [==============================] - 0s 20us/step
[0.3051770512676239, 0.87472]
