为什么在将数据馈入张量流之前将其展平？ [英] Why do we flatten the data before we feed it into tensorflow?

问题描述

我正在关注 udacity MNIST教程，而MNIST数据最初是 28 * 28 矩阵。但是，在馈入数据之前，它们将数据平整为784列（784 = 28 * 28）的1d数组。

I'm following udacity MNIST tutorial and MNIST data is originally 28*28 matrix. However right before feeding that data, they flatten the data into 1d array with 784 columns (784 = 28 * 28).

例如，
的原始训练集形状为（200000，28，28）。

200000行（数据）。每个数据都是28 * 28矩阵

For example, original training set shape was (200000, 28, 28).
200000 rows (data). Each data is 28*28 matrix

他们将其转换为形状为（200000，784）的训练集

They converted this into the training set whose shape is (200000, 784)

有人可以解释为什么他们在馈入张量流之前将数据弄平吗？

Can someone explain why they flatten the data out before feeding to tensorflow?

推荐答案

因为当您添加完全连接的图层，您始终希望数据是（1或）2维矩阵，其中每一行都是代表数据的矢量。这样，完全连接的图层就是输入（大小为（batch_size，n_features））和权重（形状为（ n_features，n_outputs））（加上偏差和激活函数），您将得到形状为（batch_size，n_outputs）的输出。另外，您真的不需要完全连接的层中的原始形状信息，因此可以将其丢失。

Because when you're adding a fully connected layer, you always want your data to be a (1 or) 2 dimensional matrix, where each row is the vector representing your data. That way, the fully connected layer is just a matrix multiplication between your input (of size (batch_size, n_features)) and the weights (of shape (n_features, n_outputs)) (plus the bias and the activation function), and you get an output of shape (batch_size, n_outputs). Plus, you really don't need the original shape information in a fully connected layer, so it's OK to lose it.

获取它会更复杂且效率更低同样的结果而无需先重塑，这就是为什么我们总是在完全连接的层之前这样做。相反，对于卷积层，您需要将数据保留为原始格式（宽度，高度）。

It would be more complicated and less efficient to get the same result without reshaping first, that's why we always do it before a fully connected layer. For a convolutional layer, on the opposite, you'll want to keep the data in original format (width, height).

这篇关于为什么在将数据馈入张量流之前将其展平？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！