Tensorflow2.0学习-基础分类 (二)

上一章是对MNIST的手写数字分类，这次对Fashion MNIST进行分类，基本都一样的，只不过换了一个数据库，还是用fit函数进行训练，fit对规整的数据集，训练还是蛮方便的。官方教程

文章目录

• 服装图像分类
• • 引包
  • 数据准备
  • 模型准备
  • 跑起来
  • 测试
  • 整体代码
• 基本文本分类
• • 引包
  • 数据准备
  • 模型准备
  • 跑起来
  • 完整代码

服装图像分类

引包

多了一个matplotlib包，就是画图用的。

# TensorFlow and tf.keras
import tensorflow as tf
from tensorflow import keras

# Helper libraries
import numpy as np
import matplotlib.pyplot as plt

数据准备

这次是Fashion MNIST，其实都一样，一个图给一个标签。图嘛，还是将其缩小至0-1之间

fashion_mnist = keras.datasets.fashion_mnist

(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()

class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',
               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']

train_images = train_images / 255.0
test_images = test_images / 255.0

模型准备

还是若干简单层的堆叠，优化器和损失函数安排上。

model = keras.Sequential([
    # 格式化数据
    keras.layers.Flatten(input_shape=(28, 28)),
    # 128个结点
    keras.layers.Dense(128, activation='relu'),
    # 10个结点
    keras.layers.Dense(10)
])

model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])

跑起来

跑10次

model.fit(train_images, train_labels, epochs=10)
# 准确率
test_loss, test_acc = model.evaluate(test_images,  test_labels, verbose=2)

测试

跑完模型可以对数据进行测试，加入Softmax将 logits 转换成更容易理解的概率。

# 对整体
probability_model = tf.keras.Sequential([model, 
                                         tf.keras.layers.Softmax()])
predictions = probability_model.predict(test_images)

# 对单个数据
# Grab an image from the test dataset.
img = test_images[1]
# Add the image to a batch where it's the only member.
img = (np.expand_dims(img,0))

predictions_single = probability_model.predict(img)
print(predictions_single)

np.argmax(predictions_single[0])

整体代码

结果

# TensorFlow and tf.keras
import tensorflow as tf
from tensorflow import keras

# Helper libraries
import numpy as np
import matplotlib.pyplot as plt

print(tf.__version__)

fashion_mnist = keras.datasets.fashion_mnist

(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()

class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',
               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']

train_images = train_images / 255.0
test_images = test_images / 255.0

model = keras.Sequential([
    keras.layers.Flatten(input_shape=(28, 28)),
    keras.layers.Dense(128, activation='relu'),
    keras.layers.Dense(10)
])

model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])

model.fit(train_images, train_labels, epochs=10)
# 准确率
test_loss, test_acc = model.evaluate(test_images,  test_labels, verbose=2)
print('\nTest accuracy:', test_acc)

Epoch 1/10
1875/1875 [==============================] - 1s 542us/step - loss: 0.5032 - accuracy: 0.8245
Epoch 2/10
1875/1875 [==============================] - 1s 544us/step - loss: 0.3774 - accuracy: 0.8630
Epoch 3/10
1875/1875 [==============================] - 1s 555us/step - loss: 0.3375 - accuracy: 0.8772
Epoch 4/10
1875/1875 [==============================] - 1s 539us/step - loss: 0.3139 - accuracy: 0.8850
Epoch 5/10
1875/1875 [==============================] - 1s 539us/step - loss: 0.2970 - accuracy: 0.8911
Epoch 6/10
1875/1875 [==============================] - 1s 541us/step - loss: 0.2819 - accuracy: 0.8957
Epoch 7/10
1875/1875 [==============================] - 1s 539us/step - loss: 0.2698 - accuracy: 0.8987
Epoch 8/10
1875/1875 [==============================] - 1s 547us/step - loss: 0.2585 - accuracy: 0.9036
Epoch 9/10
1875/1875 [==============================] - 1s 549us/step - loss: 0.2478 - accuracy: 0.9079
Epoch 10/10
1875/1875 [==============================] - 1s 545us/step - loss: 0.2391 - accuracy: 0.9112
Test accuracy: 0.8848000168800354

基本文本分类

引包

import tensorflow as tf
from tensorflow import keras

import numpy as np

print(tf.__version__)

数据准备

这次的数据是IMDB 数据集，就是电影评论，判别积极还是消极，二分类。

该数据集是经过预处理的：每个样本都是一个表示影评中词汇的整数数组。每个标签都是一个值为 0 或 1 的整数值，其中 0 代表消极评论，1代表积极评论。

文本相对于图像的挑战就是，文本都不是等长的，需要进行处理。就是按照最长的为标准，剩下的数据填充就好了，这样就有了max_length * num_reviews。

imdb = keras.datasets.imdb

(train_data, train_labels), (test_data, test_labels) = imdb.load_data(num_words=10000)
# 填充数组
train_data = keras.preprocessing.sequence.pad_sequences(train_data,
                                                        value=0,
                                                        padding='post',
                                                        maxlen=256)

test_data = keras.preprocessing.sequence.pad_sequences(test_data,
                                                       value=0,
                                                       padding='post',
                                                       maxlen=256)
# 准备验证数据
x_val = train_data[:10000]
partial_x_train = train_data[10000:]

y_val = train_labels[:10000]
partial_y_train = train_labels[10000:]

执行完代码后，train_data 的格式就变成(25000, 256)，数据规整了，就又可以送入网络了

模型准备

这里先将文本数据嵌入处理，这么做增加了维度，也能让模型效果更好一点。
然后就是池化
最后全连接

# 输入形状是用于电影评论的词汇数目（10,000 词）
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='relu'))
model.add(keras.layers.Dense(1, activation='sigmoid'))

# 打印模型
# model.summary() 

model.compile(optimizer='adam',
              loss='binary_crossentropy',
              metrics=['accuracy'])

跑起来

这里除了训练集还有验证集，验证集就是帮助我们选择哪一个模型效果最好。

verbose：日志显示
verbose = 0 为不在标准输出流输出日志信息
verbose = 1 为输出进度条记录
verbose = 2 为每个epoch输出一行记录

history = model.fit(partial_x_train,
                    partial_y_train,
                    epochs=40,
                    batch_size=512,
                    validation_data=(x_val, y_val),
                    verbose=1)
# 看下测试集                    
results = model.evaluate(test_data, test_labels, verbose=2)
print(results)

完整代码

import tensorflow as tf
from tensorflow import keras

import numpy as np

imdb = keras.datasets.imdb

(train_data, train_labels), (test_data, test_labels) = imdb.load_data(num_words=10000)

train_data = keras.preprocessing.sequence.pad_sequences(train_data,
                                                        value=0,
                                                        padding='post',
                                                        maxlen=256)

test_data = keras.preprocessing.sequence.pad_sequences(test_data,
                                                       value=0,
                                                       padding='post',
                                                       maxlen=256)

# 输入形状是用于电影评论的词汇数目（10,000 词）
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='relu'))
model.add(keras.layers.Dense(1, activation='sigmoid'))

model.summary()


model.compile(optimizer='adam',
              loss='binary_crossentropy',
              metrics=['accuracy'])

x_val = train_data[:10000]
partial_x_train = train_data[10000:]

y_val = train_labels[:10000]
partial_y_train = train_labels[10000:]

history = model.fit(partial_x_train,
                    partial_y_train,
                    epochs=40,
                    batch_size=512,
                    validation_data=(x_val, y_val),
                    verbose=1)
results = model.evaluate(test_data,  test_labels, verbose=2)

print(results)

history_dict = history.history
history_dict.keys()

import matplotlib.pyplot as plt

acc = history_dict['accuracy']
val_acc = history_dict['val_accuracy']
loss = history_dict['loss']
val_loss = history_dict['val_loss']

epochs = range(1, len(acc) + 1)

# “bo”代表 "蓝点"
plt.plot(epochs, loss, 'bo', label='Training loss')
# b代表“蓝色实线”
plt.plot(epochs, val_loss, 'b', label='Validation loss')
plt.title('Training and validation loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.legend()

plt.show()

plt.clf()   # 清除数字

plt.plot(epochs, acc, 'bo', label='Training acc')
plt.plot(epochs, val_acc, 'b', label='Validation acc')
plt.title('Training and validation accuracy')
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.legend()

plt.show()