在tensorflow中使用tf.nn.conv2d_transpose获取反卷积层的输出形状 [英] Getting the output shape of deconvolution layer using tf.nn.conv2d_transpose in tensorflow

问题描述

根据此纸张，输出形状为N + H - 1，N是输入的高度或宽度，H是内核的高度或宽度.这是明显的卷积逆过程.此教程给出了一个计算卷积输出形状的公式，该公式为(W−F+2P)/S+1，-输入大小，F-过滤器大小，P-填充大小，S-跨度.但是在 Tensorflow 中，有一些测试用例，例如:

According to this paper, the output shape is N + H - 1, N is input height or width, H is kernel height or width. This is obvious inverse process of convolution. This tutorial gives a formula to calculate the output shape of convolution which is (W−F+2P)/S+1, W - input size, F - filter size, P - padding size, S - stride. But in Tensorflow, there are test cases like:

  strides = [1, 2, 2, 1]

  # Input, output: [batch, height, width, depth]
  x_shape = [2, 6, 4, 3]
  y_shape = [2, 12, 8, 2]

  # Filter: [kernel_height, kernel_width, output_depth, input_depth]
  f_shape = [3, 3, 2, 3]

因此，根据公式(W−F+2P)/S+1，我们使用y_shape，f_shape和x_shape来计算填充大小P.从(12 - 3 + 2P) / 2 + 1 = 6，我们得到P = 0.5，它不是整数.去卷积如何在Tensorflow中工作?

So we use y_shape, f_shape and x_shape, according to formula (W−F+2P)/S+1 to calculate padding size P. From (12 - 3 + 2P) / 2 + 1 = 6, we get P = 0.5, which is not an integer. How does deconvolution works in Tensorflow?

推荐答案

用于反卷积，

output_size = strides * (input_size-1) + kernel_size - 2*padding

步幅，input_size，kernel_size，padding为整数 对于有效"，填充为零

strides, input_size, kernel_size, padding are integer padding is zero for 'valid'

这篇关于在tensorflow中使用tf.nn.conv2d_transpose获取反卷积层的输出形状的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！