Tensotflow1.0入门(五)-tensorflow的池化函数

港控/mmm° 2022-02-24 10:20 233阅读 0赞

[Tensotflow1.0入门(一)-tensorflow的设计理念和基本概念][Tensotflow1.0_-tensorflow]  
[Tensotflow1.0入门(二)-tensorflow的各种操作][Tensotflow1.0_-tensorflow 1]  
[Tensotflow1.0入门(三)-tensorflow的小例子][Tensotflow1.0_-tensorflow 2]  
[Tensotflow1.0入门(四)-tensorflow的卷积函数][Tensotflow1.0_-tensorflow 3]  
[Tensotflow1.0入门(五)-tensorflow的池化函数][Tensotflow1.0_-tensorflow 4]  
[Tensotflow1.0入门(六)-tensorflow的激活函数][Tensotflow1.0_-tensorflow 5]

池化函数的定义在tensorflow/python/ops/nn.py和gen\_nn\_ops.py(编译tensorflow后生成)中，但其实这个nn.py是由很多别的文件中的类和方法组成的

\#池化函数  
tf.nn.avg\_pool(value, ksize, strides, padding, data\_format, name)  
tf.nn.max\_pool(value, ksize, strides, padding, data\_format, name)  
tf.nn.max\_pool\_with\_argmax(input, ksize, strides, padding, Targmax, name)  
tf.nn.avg\_pool3d(input, ksize, strides, padding, name)  
tf.nn.max\_pool3d(input, ksize, strides, padding, name)  
tf.nn.fractional\_avg\_pool(value, pooling\_ratio, pseudo\_random, overlapping, deterministic, seed, seed2, name)  
tf.nn.fractional\_max\_pool(value, pooling\_ratio, pseudo\_random, overlapping, deterministic, seed, seed2, name)  
tf.nn.pool(input, window\_shape, pooling\_type, padding, dilation\_rate, strides, name, data\_format)

\#参数  
value: 4维张量，数据维度\[batch,height,width,channels\]  
ksize: 长度大于等于4的整形数组，每一位对应value中每一维  
strides: 长度大于等于4的整形数组,步长  
padding: 字符串，字符串，“SAME”|“VALID”,SAME输入数据维度和输出维度相同  
data\_format: ‘NHWC’,N个数，H为高度，W为宽度，C为通道数  
name: 操作名

例子:

import tensorflow as tf
    import numpy as np
    
    # (batch_size,w,h,channels_last)=(32,9,9,3)
    input_data = tf.Variable(np.random.rand(32,9,9,3),dtype=np.float32)
    # (w,h,channels_last,filter_nums)=(32,9,9,3)
    filter_data = tf.Variable(np.random.rand(2,2,3,5),dtype=np.float32)
    y = tf.nn.conv2d(input_data, filter_data, strides=[1,1,1,1], padding="SAME")
    out_put = tf.nn.avg_pool(y, ksize=[1,2,2,1], strides=[1,1,1,1], padding="SAME")
    init = tf.global_variables_initializer()
    sess = tf.Session()
    sess.run(init)
    print(sess.run(y).shape)
    print(sess.run(out_put).shape)
    # (32, 9, 9, 5)
    # (32, 9, 9, 5)

参考:《TensorFlow 技术解析与实战》

[Tensotflow1.0_-tensorflow]: https://blog.csdn.net/u011311291/article/details/88959590
[Tensotflow1.0_-tensorflow 1]: https://blog.csdn.net/u011311291/article/details/88959599
[Tensotflow1.0_-tensorflow 2]: https://blog.csdn.net/u011311291/article/details/89057972
[Tensotflow1.0_-tensorflow 3]: https://blog.csdn.net/u011311291/article/details/89000644
[Tensotflow1.0_-tensorflow 4]: https://blog.csdn.net/u011311291/article/details/89003525
[Tensotflow1.0_-tensorflow 5]: https://blog.csdn.net/u011311291/article/details/89000263