通过CIFAR-10数据集完成AlexNet卷积神经网络分类器的搭建

首先，我们需要导入必要的库和数据集：

import tensorflow as tf
from tensorflow.keras.datasets import cifar10

# 加载数据集
(x_train, y_train), (x_test, y_test) = cifar10.load_data()

# 归一化
x_train, x_test = x_train / 255.0, x_test / 255.0

# 转换标签为独热编码
y_train = tf.keras.utils.to_categorical(y_train, num_classes=10)
y_test = tf.keras.utils.to_categorical(y_test, num_classes=10)

接下来，我们可以开始搭建AlexNet模型。在这里，我们使用TensorFlow的Sequential模型。AlexNet共有8层，包括5个卷积层和3个全连接层。我们可以使用tf.keras.layers下的Conv2D、MaxPooling2D和Dense来构建模型：

model = tf.keras.models.Sequential([
    # 第一层卷积层
    tf.keras.layers.Conv2D(filters=96, kernel_size=(11,11), strides=(4,4), activation='relu', input_shape=(32,32,3)),
    tf.keras.layers.MaxPooling2D(pool_size=(3,3), strides=(2,2)),
    # 第二层卷积层
    tf.keras.layers.Conv2D(filters=256, kernel_size=(5,5), strides=(1,1), padding='same', activation='relu'),
    tf.keras.layers.MaxPooling2D(pool_size=(3,3), strides=(2,2)),
    # 第三层卷积层
    tf.keras.layers.Conv2D(filters=384, kernel_size=(3,3), strides=(1,1), padding='same', activation='relu'),
    # 第四层卷积层
    tf.keras.layers.Conv2D(filters=384, kernel_size=(3,3), strides=(1,1), padding='same', activation='relu'),
    # 第五层卷积层
    tf.keras.layers.Conv2D(filters=256, kernel_size=(3,3), strides=(1,1), padding='same', activation='relu'),
    tf.keras.layers.MaxPooling2D(pool_size=(3,3), strides=(2,2)),
    # 全连接层
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(4096, activation='relu'),
    tf.keras.layers.Dropout(0.5),
    tf.keras.layers.Dense(4096, activation='relu'),
    tf.keras.layers.Dropout(0.5),
    tf.keras.layers.Dense(10, activation='softmax')
])

在这里，我们使用ReLU激活函数，并在全连接层中使用Dropout以避免过拟合。

接下来，我们可以编译模型并开始训练：

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

model.fit(x_train, y_train, epochs=10, batch_size=128, validation_data=(x_test, y_test))

最后，我们可以使用测试集来评估模型的准确率：

test_loss, test_acc = model.evaluate(x_test, y_test)
print('Test accuracy:', test_acc)

完整代码如下：

import tensorflow as tf
from tensorflow.keras.datasets import cifar10

# 加载数据集
(x_train, y_train), (x_test, y_test) = cifar10.load_data()

# 归一化
x_train, x_test = x_train / 255.0, x_test / 255.0

# 转换标签为独热编码
y_train = tf.keras.utils.to_categorical(y_train, num_classes=10)
y_test = tf.keras.utils.to_categorical(y_test, num_classes=10)

model = tf.keras.models.Sequential([
    # 第一层卷积层
    tf.keras.layers.Conv2D(filters=96, kernel_size=(11,11), strides=(4,4), activation='relu', input_shape=(32,32,3)),
    tf.keras.layers.MaxPooling2D(pool_size=(3,3), strides=(2,2)),
    # 第二层卷积层
    tf.keras.layers.Conv2D(filters=256, kernel_size=(5,5), strides=(1,1), padding='same', activation='relu'),
    tf.keras.layers.MaxPooling2D(pool_size=(3,3), strides=(2,2)),
    # 第三层卷积层
    tf.keras.layers.Conv2D(filters=384, kernel_size=(3,3), strides=(1,1), padding='same', activation='relu'),
    # 第四层卷积层
    tf.keras.layers.Conv2D(filters=384, kernel_size=(3,3), strides=(1,1), padding='same', activation='relu'),
    # 第五层卷积层
    tf.keras.layers.Conv2D(filters=256, kernel_size=(3,3), strides=(1,1), padding='same', activation='relu'),
    tf.keras.layers.MaxPooling2D(pool_size=(3,3), strides=(2,2)),
    # 全连接层
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(4096, activation='relu'),
    tf.keras.layers.Dropout(0.5),
    tf.keras.layers.Dense(4096, activation='relu'),
    tf.keras.layers.Dropout(0.5),
    tf.keras.layers.Dense(10, activation='softmax')
])

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

model.fit(x_train, y_train, epochs=10, batch_size=128, validation_data=(x_test, y_test))

test_loss, test_acc = model.evaluate(x_test, y_test)
print('Test accuracy:', test_acc)