实现一个简单的网络训练---tensorflow2.0学习

处理数据
dataset = tf.data.Dataset.from_tensor_slices((train_x, train_y))
dataset = dataset.shuffle(4)#打乱，越大越乱
dataset = dataset.batch(32)#batch——size
dataset = dataset.repeat()#数据集重复次数，一般默认为空,其实就是epoch
#创建一个迭代器
#iterator = dataset.make_one_shot_iterator（）
#dataset = dataset.get_next()

构造模型：
model = tf.keras.Sequential()
model.add(layers.Dense(units,activation…))
…
model.add(layers.Dense())
配置模型
model.compile(optimizer=tf.keras.optimizers.Adam(0.001),
loss=tf.keras.losses.categorical_crossentropy,
metrics=[tf.keras.metrics.categorical_accuracy])
开始训练
history = model.fit(x_train, y_train, batch_size=64, epochs=5,validation_split=0.1)
画图
plt.plot(history.history[‘accuracy’])
plt.plot(history.history[‘val_accuracy’])
plt.legend([‘training’, ‘valivation’], loc=‘upper left’)
plt.show()

model.fit函数的详解：https://blog.csdn.net/a1111h/article/details/82148497
评估
model.evaluate(test_x, test_y, batch_size=32)
预测
result = model.predict(test_x, batch_size=32)

完整代码：

```python
from tensorflow import keras
from tensorflow.keras import layers
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

x_train = np.random.random(1000)
#noise = np.random.random(1000) * 0.1
y_train = np.random.randint(0,2,1000)

#创建模型
model = keras.Sequential()
model.add(layers.Dense(32))
model.add(layers.Dense(1,activation='sigmoid'))
#配置模型
model.compile(optimizer=tf.keras.optimizers.Adam(0.01),
			  loss=keras.losses.BinaryCrossentropy(),
			  metrics=['accuracy'])

#训练模型
history = model.fit(x_train,y_train,epochs=5)
#画图
model.summary()

plt.plot(history.history['accuracy'])
#plt.plot(history.history['val_accuracy'])
#plt.legend(['training', 'valivation'], loc='upper left') 
plt.show()
#评估模型

#预测数据
print(model.predict([1,2,3]))