【深度学习】常见数据集和常见卷积神经网络

MNIST 数据集

class MNISTLoader():
    def __init__(self):
        mnist = tf.keras.datasets.mnist
        (self.train_data, self.train_label), (self.test_data, self.test_label) = mnist.load_data()
        self.test=self.test_data.astype(np.float32)
        # MNIST中的图像默认为uint8（0-255的数字）。以下代码将其归一化到0-1之间的浮点数，并在最后增加一维作为颜色通道
        self.train_data = np.expand_dims(self.train_data.astype(np.float32) / 255.0, axis=-1)      # [60000, 28, 28, 1]
        self.test_data = np.expand_dims(self.test_data.astype(np.float32) / 255.0, axis=-1)        # [10000, 28, 28, 1]
        self.train_label = self.train_label.astype(np.int32)    # [60000]
        self.test_label = self.test_label.astype(np.int32)      # [10000]
        self.num_train_data, self.num_test_data = self.train_data.shape[0], self.test_data.shape[0]

    def get_batch(self, batch_size):
        # 从数据集中随机取出batch_size个元素并返回
        index = np.random.randint(0, self.num_train_data, batch_size)
        return self.train_data[index, :], self.train_label[index]

Cifar10 数据集

class Cifar10Loader():
    def __init__(self):
        cifar10 = tf.keras.datasets.cifar10
        (self.train_data, self.train_label), (self.test_data, self.test_label) = cifar10.load_data()
        self.test=self.test_data.astype(np.float32)
        # MNIST中的图像默认为uint8（0-255的数字）。以下代码将其归一化到0-1之间的浮点数，并在最后增加一维作为颜色通道
        self.train_data = self.train_data.astype(np.float32) / 255.0   # [50000, 32, 32, 3]
        self.test_data = self.test_data.astype(np.float32) / 255.0     # [10000, 32, 32, 3]
        self.train_label = np.squeeze(self.train_label.astype(np.int32))    # [50000]
        self.test_label = np.squeeze(self.test_label.astype(np.int32))      # [10000]
        self.num_train_data, self.num_test_data = self.train_data.shape[0], self.test_data.shape[0]

    def get_batch(self, batch_size):
        # 从数据集中随机取出batch_size个元素并返回
        index = np.random.randint(0, self.num_train_data, batch_size)
        return self.train_data[index, :], self.train_label[index]

MLP 全连通网络

class MLP(tf.keras.Model): 
    def __init__(self):
        super().__init__()
        self.flatten = tf.keras.layers.Flatten()    # Flatten层将除第一维（batch_size）以外的维度展平
        self.dense1 = tf.keras.layers.Dense(units=100, activation=tf.nn.relu)
        self.dense2 = tf.keras.layers.Dense(units=10)

    def call(self, inputs):         # [batch_size, 28, 28, 1]
        x = self.flatten(inputs)    # [batch_size, 784]
        x = self.dense1(x)          # [batch_size, 100]
        x = self.dense2(x)          # [batch_size, 10]
        output = tf.nn.softmax(x)
        return output

CNN 卷积神经网络

class CNN(tf.keras.Model):
    def __init__(self):
        super().__init__()
        self.conv1 = tf.keras.layers.Conv2D(
            filters=32,             # 卷积层神经元（卷积核）数目
            kernel_size=[5, 5],     # 感受野大小
            padding='same',         # padding策略（vaild 或 same）
            activation=tf.nn.relu   # 激活函数
        )
        self.pool1 = tf.keras.layers.MaxPool2D(pool_size=[2, 2], strides=2)
        self.conv2 = tf.keras.layers.Conv2D(
            filters=64,
            kernel_size=[5, 5],
            padding='same',
            activation=tf.nn.relu
        )
        self.pool2 = tf.keras.layers.MaxPool2D(pool_size=[2, 2], strides=2)
        self.flatten = tf.keras.layers.Reshape(target_shape=(8 * 8  * 64,))
        self.dense1 = tf.keras.layers.Dense(units=1024, activation=tf.nn.relu)
        self.dense2 = tf.keras.layers.Dense(units=10)

    def call(self, inputs):
        x = self.conv1(inputs)                  # [batch_size, 32, 32, 32]
        x = self.pool1(x)                       # [batch_size, 16, 16, 32]
        x = self.conv2(x)                       # [batch_size, 16, 16, 64]
        x = self.pool2(x)                       # [batch_size, 8, 8, 64]
        x = self.flatten(x)                     # [batch_size, 8 * 8 * 64]
        x = self.dense1(x)                      # [batch_size, 1024]
        x = self.dense2(x)                      # [batch_size, 10]
        output = tf.nn.softmax(x)
        return output

识别手写数字

训练网络

import numpy as np
import tensorflow as tf
from matplotlib import pyplot as plt

import os
os.environ['CUDA_VISIBLE_DEVICES'] = '0'

class MNISTLoader():
    def __init__(self):
        mnist = tf.keras.datasets.mnist
        (self.train_data, self.train_label), (self.test_data, self.test_label) = mnist.load_data()
        self.test=self.test_data.astype(np.float32)
        # MNIST中的图像默认为uint8（0-255的数字）。以下代码将其归一化到0-1之间的浮点数，并在最后增加一维作为颜色通道
        self.train_data = np.expand_dims(self.train_data.astype(np.float32) / 255.0, axis=-1)      # [60000, 28, 28, 1]
        self.test_data = np.expand_dims(self.test_data.astype(np.float32) / 255.0, axis=-1)        # [10000, 28, 28, 1]
        self.train_label = self.train_label.astype(np.int32)    # [60000]
        self.test_label = self.test_label.astype(np.int32)      # [10000]
        self.num_train_data, self.num_test_data = self.train_data.shape[0], self.test_data.shape[0]

    def get_batch(self, batch_size):
        # 从数据集中随机取出batch_size个元素并返回
        index = np.random.randint(0, self.num_train_data, batch_size)
        return self.train_data[index, :], self.train_label[index]

    
class CNN(tf.keras.Model):
    def __init__(self):
        super().__init__()
        self.conv1 = tf.keras.layers.Conv2D(
            filters=32,             # 卷积层神经元（卷积核）数目
            kernel_size=[5, 5],     # 感受野大小
            padding='same',         # padding策略（vaild 或 same）
            activation=tf.nn.relu   # 激活函数
        )
        self.pool1 = tf.keras.layers.MaxPool2D(pool_size=[2, 2], strides=2)
        self.conv2 = tf.keras.layers.Conv2D(
            filters=64,
            kernel_size=[5, 5],
            padding='same',
            activation=tf.nn.relu
        )
        self.pool2 = tf.keras.layers.MaxPool2D(pool_size=[2, 2], strides=2)
        self.flatten = tf.keras.layers.Reshape(target_shape=(7 * 7 * 64,))
        self.dense1 = tf.keras.layers.Dense(units=1024, activation=tf.nn.relu)
        self.dense2 = tf.keras.layers.Dense(units=10)

    def call(self, inputs):
        x = self.conv1(inputs)                  # [batch_size, 28, 28, 32]
        x = self.pool1(x)                       # [batch_size, 14, 14, 32]
        x = self.conv2(x)                       # [batch_size, 14, 14, 64]
        x = self.pool2(x)                       # [batch_size, 7, 7, 64]
        x = self.flatten(x)                     # [batch_size, 7 * 7 * 64]
        x = self.dense1(x)                      # [batch_size, 1024]
        x = self.dense2(x)                      # [batch_size, 10]
        output = tf.nn.softmax(x)
        return output


num_epochs = 1
batch_size = 20
learning_rate = 0.003

model = CNN()
data_loader = MNISTLoader()
model.compile(
    optimizer=tf.keras.optimizers.Adam(learning_rate=0.003),
    loss=tf.keras.losses.sparse_categorical_crossentropy,
    metrics=[tf.keras.metrics.sparse_categorical_accuracy]
)
model.fit(data_loader.train_data, data_loader.train_label, epochs=num_epochs, batch_size=batch_size)
tf.saved_model.save(model, "saved")

验证测试集

import tensorflow as tf
import numpy as np

import os
os.environ['CUDA_VISIBLE_DEVICES'] = '-1'

class MNISTLoader():
    def __init__(self):
        mnist = tf.keras.datasets.mnist
        (self.train_data, self.train_label), (self.test_data, self.test_label) = mnist.load_data()
        self.test=self.test_data.astype(np.float32)
        # MNIST中的图像默认为uint8（0-255的数字）。以下代码将其归一化到0-1之间的浮点数，并在最后增加一维作为颜色通道
        self.train_data = np.expand_dims(self.train_data.astype(np.float32) / 255.0, axis=-1)      # [60000, 28, 28, 1]
        self.test_data = np.expand_dims(self.test_data.astype(np.float32) / 255.0, axis=-1)        # [10000, 28, 28, 1]
        self.train_label = self.train_label.astype(np.int32)    # [60000]
        self.test_label = self.test_label.astype(np.int32)      # [10000]
        self.num_train_data, self.num_test_data = self.train_data.shape[0], self.test_data.shape[0]

    def get_batch(self, batch_size):
        # 从数据集中随机取出batch_size个元素并返回
        index = np.random.randint(0, self.num_train_data, batch_size)
        return self.train_data[index, :], self.train_label[index]
    


batch_size = 50

model = tf.saved_model.load("saved")
data_loader = MNISTLoader()
sparse_categorical_accuracy = tf.keras.metrics.SparseCategoricalAccuracy()
num_batches = int(data_loader.num_test_data // batch_size)
for batch_index in range(num_batches):
    start_index, end_index = batch_index * batch_size, (batch_index + 1) * batch_size
    y_pred = model(data_loader.test_data[start_index: end_index])
    sparse_categorical_accuracy.update_state(y_true=data_loader.test_label[start_index: end_index], y_pred=y_pred)
print("test accuracy: %f" % sparse_categorical_accuracy.result())

Cifar10-CNN

训练网络

import numpy as np
import tensorflow as tf
from matplotlib import pyplot as plt
#from TensorFlow import Cifar10Loader, CNN

import os
os.environ['CUDA_VISIBLE_DEVICES'] = '0'

class CNN(tf.keras.Model):
    def __init__(self):
        super().__init__()
        self.conv1 = tf.keras.layers.Conv2D(
            filters=32,             # 卷积层神经元（卷积核）数目
            kernel_size=[5, 5],     # 感受野大小
            padding='same',         # padding策略（vaild 或 same）
            activation=tf.nn.relu   # 激活函数
        )
        self.pool1 = tf.keras.layers.MaxPool2D(pool_size=[2, 2], strides=2)
        self.conv2 = tf.keras.layers.Conv2D(
            filters=64,
            kernel_size=[5, 5],
            padding='same',
            activation=tf.nn.relu
        )
        self.pool2 = tf.keras.layers.MaxPool2D(pool_size=[2, 2], strides=2)
        self.flatten = tf.keras.layers.Reshape(target_shape=(8 * 8  * 64,))
        self.dense1 = tf.keras.layers.Dense(units=1024, activation=tf.nn.relu)
        self.dense2 = tf.keras.layers.Dense(units=10)

    def call(self, inputs):
        x = self.conv1(inputs)                  # [batch_size, 32, 32, 32]
        x = self.pool1(x)                       # [batch_size, 16, 16, 32]
        x = self.conv2(x)                       # [batch_size, 16, 16, 64]
        x = self.pool2(x)                       # [batch_size, 8, 8, 64]
        x = self.flatten(x)                     # [batch_size, 8 * 8 * 64]
        x = self.dense1(x)                      # [batch_size, 1024]
        x = self.dense2(x)                      # [batch_size, 10]
        output = tf.nn.softmax(x)
        return output

class Cifar10Loader():
    def __init__(self):
        cifar10 = tf.keras.datasets.cifar10
        (self.train_data, self.train_label), (self.test_data, self.test_label) = cifar10.load_data()
        self.test=self.test_data.astype(np.float32)
        # MNIST中的图像默认为uint8（0-255的数字）。以下代码将其归一化到0-1之间的浮点数，并在最后增加一维作为颜色通道
        self.train_data = self.train_data.astype(np.float32) / 255.0   # [50000, 32, 32, 3]
        self.test_data = self.test_data.astype(np.float32) / 255.0     # [10000, 32, 32, 3]
        self.train_label = np.squeeze(self.train_label.astype(np.int32))    # [50000]
        self.test_label = np.squeeze(self.test_label.astype(np.int32))      # [10000]
        self.num_train_data, self.num_test_data = self.train_data.shape[0], self.test_data.shape[0]

    def get_batch(self, batch_size):
        # 从数据集中随机取出batch_size个元素并返回
        index = np.random.randint(0, self.num_train_data, batch_size)
        return self.train_data[index, :], self.train_label[index]


num_epochs = 1
batch_size = 50
learning_rate = 0.0003

model = CNN()
data_loader = Cifar10Loader()
model.compile(
    optimizer=tf.keras.optimizers.Adam(learning_rate=0.0003),
    loss=tf.keras.losses.sparse_categorical_crossentropy,
    metrics=[tf.keras.metrics.sparse_categorical_accuracy]
)
model.fit(data_loader.train_data, data_loader.train_label, epochs=num_epochs, batch_size=batch_size)
tf.saved_model.save(model, "saved")

验证测试集

import tensorflow as tf
import numpy as np
from TensorFlow import Cifar10Loader, CNN

import os
os.environ['CUDA_VISIBLE_DEVICES'] = '-1'

batch_size = 50

model = tf.saved_model.load("saved")
data_loader = Cifar10Loader()
sparse_categorical_accuracy = tf.keras.metrics.SparseCategoricalAccuracy()
num_batches = int(data_loader.num_test_data // batch_size)
for batch_index in range(num_batches):
    start_index, end_index = batch_index * batch_size, (batch_index + 1) * batch_size
    y_pred = model(data_loader.test_data[start_index: end_index])
    sparse_categorical_accuracy.update_state(y_true=data_loader.test_label[start_index: end_index], y_pred=y_pred)
print("test accuracy: %f" % sparse_categorical_accuracy.result())