TensorFlow2教程-基础MLP网络

版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/qq_31456593/article/details/88915936

Tensorflow 2.0 教程持续更新 ：https://blog.csdn.net/qq_31456593/article/details/88606284

完整tensorflow2.0教程代码请看tensorflow2.0：中文教程tensorflow2_tutorials_chinese(欢迎star)

入门教程：
TensorFlow 2.0 教程- Keras 快速入门
TensorFlow 2.0 教程-keras 函数api
TensorFlow 2.0 教程-使用keras训练模型
TensorFlow 2.0 教程-用keras构建自己的网络层
TensorFlow 2.0 教程-keras模型保存和序列化

TensorFlow2教程-基础MLP网络

1.回归任务

# 导入数据
(x_train, y_train), (x_test, y_test) = keras.datasets.boston_housing.load_data()
print(x_train.shape, ' ', y_train.shape)
print(x_test.shape, ' ', y_test.shape)

(404, 13)   (404,)
(102, 13)   (102,)

# 构建模型

model = keras.Sequential([
    layers.Dense(32, activation='sigmoid', input_shape=(13,)),
    layers.Dense(32, activation='sigmoid'),
    layers.Dense(32, activation='sigmoid'),
    layers.Dense(1)
])

# 配置模型
model.compile(optimizer=keras.optimizers.SGD(0.1),
             loss='mean_squared_error',  # keras.losses.mean_squared_error
             metrics=['mse'])
model.summary()

Model: "sequential_10"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense_33 (Dense)             (None, 32)                448       
_________________________________________________________________
dense_34 (Dense)             (None, 32)                1056      
_________________________________________________________________
dense_35 (Dense)             (None, 32)                1056      
_________________________________________________________________
dense_36 (Dense)             (None, 1)                 33        
=================================================================
Total params: 2,593
Trainable params: 2,593
Non-trainable params: 0
_________________________________________________________________

# 训练
model.fit(x_train, y_train, batch_size=50, epochs=50, validation_split=0.1, verbose=1)

Train on 363 samples, validate on 41 samples
Epoch 1/50
363/363 [==============================] - 0s 430us/sample - loss: 371.0176 - mse: 371.0175 - val_loss: 50.0381 - val_mse: 50.0381

Epoch 50/50
363/363 [==============================] - 0s 28us/sample - loss: 80.1490 - mse: 80.1490 - val_loss: 30.6706 - val_mse: 30.6706

result = model.evaluate(x_test, y_test)

102/102 [==============================] - 0s 116us/sample - loss: 75.0492 - mse: 75.0492

print(model.metrics_names)
print(result)

['loss', 'mse']
[75.04923741957721, 75.04924]

2.分类任务

from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split

whole_data = load_breast_cancer()
x_data = whole_data.data
y_data = whole_data.target

x_train, x_test, y_train, y_test = train_test_split(x_data, y_data, test_size=0.3, random_state=7)

print(x_train.shape, ' ', y_train.shape)
print(x_test.shape, ' ', y_test.shape)

(398, 30)   (398,)
(171, 30)   (171,)

# 构建模型
model = keras.Sequential([
    layers.Dense(32, activation='relu', input_shape=(30,)),
    layers.Dense(32, activation='relu'),
    layers.Dense(1, activation='sigmoid')
])

model.compile(optimizer=keras.optimizers.Adam(),
             loss=keras.losses.binary_crossentropy,
             metrics=['accuracy'])
model.summary()

Model: "sequential_14"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense_46 (Dense)             (None, 32)                992       
_________________________________________________________________
dense_47 (Dense)             (None, 32)                1056      
_________________________________________________________________
dense_48 (Dense)             (None, 1)                 33        
=================================================================
Total params: 2,081
Trainable params: 2,081
Non-trainable params: 0
_________________________________________________________________

model.fit(x_train, y_train, batch_size=64, epochs=10, verbose=1)

Epoch 10/10
398/398 [==============================] - 0s 27us/sample - loss: 0.2597 - accuracy: 0.9045

model.evaluate(x_test, y_test)

171/171 [==============================] - 0s 463us/sample - loss: 0.3877 - accuracy: 0.8772





[0.38765583248340596, 0.877193]

print(model.metrics_names)

['loss', 'accuracy']