在 TensorFlow 训练期间打印损失 [英] Printing the loss during TensorFlow training

问题描述

我正在研究 TensorFlow "MNIST For ML Beginners"教程，我想在每个训练步骤后打印出训练损失.

I am looking at the TensorFlow "MNIST For ML Beginners" tutorial, and I want to print out the training loss after every training step.

我的训练循环目前如下所示:

My training loop currently looks like this:

for i in range(100):
    batch_xs, batch_ys = mnist.train.next_batch(100)
    sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})

现在，train_step 定义为:

train_step = tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)

其中 cross_entropy 是我想打印的损失:

Where cross_entropy is the loss which I want to print out:

cross_entropy = -tf.reduce_sum(y_ * tf.log(y))

打印这个的一种方法是在训练循环中显式计算cross_entropy:

One way to print this would be to explicitly compute cross_entropy in the training loop:

for i in range(100):
    batch_xs, batch_ys = mnist.train.next_batch(100)
    cross_entropy = -tf.reduce_sum(y_ * tf.log(y))
    print 'loss = ' + str(cross_entropy)
    sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})

我现在有两个关于此的问题:

I now have two questions regarding this:

1. 鉴于 cross_entropy 已经在 sess.run(train_step, ...) 期间计算过，计算两次似乎效率低下，需要两倍的数字所有训练数据的前向传递.在 sess.run(train_step, ...) 期间计算时，有没有办法访问 cross_entropy 的值?

1. Given that cross_entropy is already computed during sess.run(train_step, ...), it seems inefficient to compute it twice, requiring twice the number of forward passes of all the training data. Is there a way to access the value of cross_entropy when it was computed during sess.run(train_step, ...)?

我什至如何打印tf.Variable?使用 str(cross_entropy) 给我一个错误...

How do I even print a tf.Variable? Using str(cross_entropy) gives me an error...

谢谢！

推荐答案

您可以通过将 cross_entropy 的值添加到 sess.run(...).例如，您的 for 循环可以重写如下:

You can fetch the value of cross_entropy by adding it to the list of arguments to sess.run(...). For example, your for-loop could be rewritten as follows:

for i in range(100):
    batch_xs, batch_ys = mnist.train.next_batch(100)
    cross_entropy = -tf.reduce_sum(y_ * tf.log(y))
    _, loss_val = sess.run([train_step, cross_entropy],
                           feed_dict={x: batch_xs, y_: batch_ys})
    print 'loss = ' + loss_val

同样的方法可用于打印变量的当前值.假设，除了 cross_entropy 的值之外，您还想打印名为 W 的 tf.Variable 的值，您可以执行以下:

The same approach can be used to print the current value of a variable. Let's say, in addition to the value of cross_entropy, you wanted to print the value of a tf.Variable called W, you could do the following:

for i in range(100):
    batch_xs, batch_ys = mnist.train.next_batch(100)
    cross_entropy = -tf.reduce_sum(y_ * tf.log(y))
    _, loss_val, W_val = sess.run([train_step, cross_entropy, W],
                                  feed_dict={x: batch_xs, y_: batch_ys})
    print 'loss = %s' % loss_val
    print 'W = %s' % W_val

这篇关于在 TensorFlow 训练期间打印损失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！