Tensorflow minibatch 训练 [英] Tensorflow minibatch training

问题描述

如何使用小批量数据在 TensorFlow 中训练网络?在 Deep-MNIST 教程中，他们使用:

How can I train a network in TensorFlow using minibatches of data? In the Deep-MNIST tutorial, they use:

for i in range(1000):
   batch = mnist.train.next_batch(50)
   train_step.run(feed_dict={x: batch[0], y_: batch[1]})

我的问题是 - x 和 y_ 变量的维度是否适合单个示例，以及 batch[0],batch[1] 是这些输入和输出的列表吗?在这种情况下，TensorFlow 会自动为这些列表中的每个训练示例添加梯度吗?或者我应该创建我的模型以便 x 和 y_ 获得整个 minibatch?

My question is - are x and y_ variables with dimensions suitable to a single example, and batch[0],batch[1] are lists of such inputs and outputs? in this case, will TensorFlow automatically add the gradients for each training example in these lists? or should I create my model so that x and y_ get an entire minibatch?

我的问题是，当我尝试为每个占位符提供一个列表时，它会尝试输入占位符的整个列表，因此我得到大小不匹配:Cannot feed value of shape (n, m) 对于 Tensor u'ts:0'，其形状为 '(m,)'，其中 n 是小批量大小，m 是单个输入尺寸.

My problem is that when I try to feed it a list for each placeholder, it tries to input the entire list for the placeholder, and I therefore get a size mismatch: Cannot feed value of shape (n, m) for Tensor u'ts:0', which has shape '(m,)', where n is the minibatch size and m is the individual input size.

谢谢.

推荐答案

在 MNIST 教程中 x 和 y_ 是具有定义形状的占位符:

In the MNIST tutorial x and y_are placeholders with a defined shape:

x = tf.placeholder(tf.float32, shape=[None, 784])
y_ = tf.placeholder(tf.float32, shape=[None, 10])

shape=[None, 784] 表示这个占位符有二维.

The shape=[None, 784] means that this placeholder have 2 dimension.

所以，回答你的第一个问题:

So, to answer your first question:

是具有适合单个示例的维度的 x 和 y_ 变量

are x and y_ variables with dimensions suitable to a single example

第一个维度可以包含未定义数量的元素(so, 1, 2, ... 50 ...)，第二个维度可以包含 784 = 28*28 个元素(这是单个 MNIST 的特征图片).

The first dimension can contain an undefined number of elements (so, 1, 2, ... 50 ...) and the second dimension can contain exaclly 784 = 28*28 elements (that are the features of a single MNIST image).

如果您使用形状为 [1, 784] 或 [50, 784] 的 Python 列表来提供图形，则 tensorflow 完全相同，它可以毫无问题地处理它.

If you feed the graph with a python list with shape [1, 784] or [50, 784] is totally the same for tensorflow, it can handle it without any problem.

batch[0],batch[1] 是这些输入和输出的列表吗?在教程中，他们定义了批处理调用batch = datasets.train.next_batch(50).因此:

batch[0],batch[1] are lists of such inputs and outputs? in the tutorial they define batch calling batch = datasets.train.next_batch(50). Thus:

• batch[0] 是一个形状为 [50, 784] 的列表
• batch[1] 是一个形状为 [50, 10] 的列表

TensorFlow 会自动为这些列表中的每个训练示例添加梯度吗?或者我应该创建我的模型，以便 x 和 y_ 获得整个小批量?

will TensorFlow automatically add the gradients for each training example in these lists? or should I create my model so that x and y_ get an entire minibatch?

Tensorflow 会为您处理这个问题.

Tensorflow handles this for you.

您报告的错误 无法为 Tensor u'ts:0' 提供形状 (n, m) 的值，其形状为 '(m,)'是形状不匹配错误.

The error you're reporting Cannot feed value of shape (n, m) for Tensor u'ts:0', which has shape '(m,)' is a shape mismatch error.

您并没有将输入重塑为具有与占位符相同的形状.

You're not reshaping the inputs to have the same shape of the placeholder.

这篇关于Tensorflow minibatch 训练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！