tf.keras.layers.Dropout 示例

import tensorflow as tf
import numpy as np

随机删除神经元，防止过度拟合

tf.random.set_seed(0)
layer = tf.keras.layers.Dropout(
    .2,  # 0~1之间的小数。要丢弃的输入数据的比例。
    input_shape=(2, ))
print(layer)

<tensorflow.python.keras.layers.core.Dropout object at 0x0000021CF0A03820>

inputs = np.arange(10).reshape(5, 2).astype(np.float32)
print(inputs)

[[0. 1.]
 [2. 3.]
 [4. 5.]
 [6. 7.]
 [8. 9.]]

outputs = layer(inputs, training=True)
print(outputs)

tf.Tensor(
[[ 0.    1.25]
 [ 2.5   3.75]
 [ 5.    6.25]
 [ 7.5   8.75]
 [10.    0.  ]], shape=(5, 2), dtype=float32)