本文共 4405 字，大约阅读时间需要 14 分钟。
Keras 的便捷性在于其代码简洁，以下是实现线性回归的示例：
import numpy as npfrom keras.models import Sequentialfrom keras.layers import Dense# 生成随机点x_data = np.random.rand(100)# 添加噪声noise = np.random.normal(0, 0.01, x_data.shape)y_data = x_data * 0.1 + 0.2 + noise# 定义模型model = Sequential()model.add(Dense(units=1, input_dim=1))# 编译模型model.compile(optimizer='sgd', loss='mse')# 训练过程for step in range(3000): cost = model.train_on_batch(x_data, y_data) if step % 300 == 0: print(f'step: {step}; cost: {cost}') # 预测结果y_pred = model.predict(x_data) 通过添加激活函数，可以实现非线性回归：
import numpy as npfrom keras.models import Sequentialfrom keras.layers import Dense, Activationfrom keras.optimizers import SGD# 随机生成200个点x_data = np.linspace(-0.5, 0.5, 200)# 添加噪声noise = np.random.normal(0, 0.02, x_data.shape)y_data = np.square(x_data) + noise# 定义模型model = Sequential()model.add(Dense(units=10, input_dim=1))model.add(Activation('tanh'))model.add(Dense(units=1, activation='tanh'))# 编译模型model.compile(optimizer=SGD(lr=0.1), loss='mse')# 训练过程for step in range(6000): cost = model.train_on_batch(x_data, y_data) if step % 300 == 0: print(f'step: {step}; cost: {cost}') 使用 Keras 实现 MNIST 的手写数字分类：
from keras.datasets import mnistfrom keras.models import Sequentialfrom keras.layers import Dense, Dropoutfrom keras.utils import np_utilsfrom keras.optimizers import SGDfrom keras.regularizers import l2# 加载数据集(x_train, y_train), (x_test, y_test) = mnist.load_data()# 数据预处理x_train = x_train.reshape(x_train.shape[0], -1) / 255x_test = x_test.reshape(x_test.shape[0], -1) / 255# 将标签转换为 one-hot 编码y_train = np_utils.to_categorical(y_train, num_classes=10)y_test = np_utils.to_categorical(y_test, num_classes=10)# 定义模型model = Sequential([ Dense(units=200, input_dim=784, bias_initializer='one', activation='relu', kernel_regularizer=l2(0.0003)), Dropout(0.4), Dense(units=100, bias_initializer='one', activation='relu'), Dropout(0.4), Dense(units=10, activation='softmax')])# 编译模型adam = SGD(lr=0.2)model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])# 训练模型model.fit(x_train, y_train, batch_size=32, epochs=10)
使用卷积层实现图像分类：
from keras.datasets import mnistfrom keras.models import Sequentialfrom keras.layers import Dense, Convolution2D, MaxPool2D, Flatten, Dropoutfrom keras.utils import np_utilsfrom keras.optimizers import Adam# 加载数据集(x_train, y_train), (x_test, y_test) = mnist.load_data()# 数据预处理x_train = x_train.reshape(-1, 28, 28, 1) / 255x_test = x_test.reshape(-1, 28, 28, 1) / 255# 将标签转换为 one-hot 编码y_train = np_utils.to_categorical(y_train, num_classes=10)y_test = np_utils.to_categorical(y_test, num_classes=10)# 定义模型model = Sequential([ Convolution2D(filters=32, kernel_size=5, strides=1, padding='same', activation='relu'), MaxPool2D(pool_size=2, strides=2, padding='same'), Convolution2D(filters=64, kernel_size=5, strides=1, padding='same', activation='relu'), MaxPool2D(pool_size=2, strides=2, padding='same'), Flatten(), Dense(1024, activation='relu'), Dropout(0.5), Dense(10, activation='softmax')])# 编译模型adam = Adam(lr=0.001)model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])# 训练模型model.fit(x_train, y_train, batch_size=64, epochs=4)
实现时间序列预测：
from keras.datasets import mnistfrom keras.models import Sequentialfrom keras.layers import Dense, LSTMfrom keras.utils import np_utilsfrom keras.optimizers import Adam# 加载数据集(x_train, y_train), (x_test, y_test) = mnist.load_data()# 数据预处理time_steps = 28input_size = 28cell_size = 50x_train = x_train / 255x_test = x_test / 255# 将标签转换为 one-hot 编码y_train = np_utils.to_categorical(y_train, num_classes=10)y_test = np_utils.to_categorical(y_test, num_classes=10)# 定义模型model = Sequential([ LSTM(units=cell_size, input_shape=(time_steps, input_size)), Dense(10, activation='softmax')])# 编译模型adam = Adam(lr=0.001)model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])# 训练模型model.fit(x_train, y_train, batch_size=64, epochs=10)
使用 h5py 库保存模型：
model.save('lstm_mnist.h5') 在需要时加载模型：
from keras.models import load_modelmodel = load_model('lstm_mnist.h5') 以上代码示例展示了 Keras 在多种任务中的实际应用，涵盖了线性回归、非线性回归、图像分类以及时间序列预测等。通过这些示例，可以清晰地看到 Keras 在机器学习模型开发中的强大优势。
转载地址：http://tzee.baihongyu.com/