Tensorflow:为什么 tf.case 给我错误的结果? [英] Tensorflow: Why is tf.case giving me the wrong result?

问题描述

我正在尝试使用 tf.case (https://www.tensorflow.org/api_docs/python/tf/case) 有条件地更新张量.如图所示，当 global_step == 2 时，我试图将 learning_rate 更新为 0.01，当 0.001global_step == 4.

I'm trying to use tf.case (https://www.tensorflow.org/api_docs/python/tf/case) to conditionally update a Tensor. As shown, I'm trying to update learning_rate to 0.01 when global_step == 2, and to 0.001 when global_step == 4.

然而，当 global_step == 2 时，我已经得到 learning_rate = 0.001.经过进一步检查，当 global_step == 2(我得到 0.001 而不是 >0.01).即使 0.01 的谓词求值为 True，0.001 的谓词求值为 False，也会发生这种情况.

However, when global_step == 2, I already get learning_rate = 0.001. Upon further inspection, it looks like tf.case is giving me the wrong result when global_step == 2 (I get 0.001 instead of 0.01). This is happening even though the predicate for 0.01 is evaluating to True, and the predicate for 0.001 is evaluating to False.

我做错了什么，还是这是一个错误?

Am I doing something wrong, or is this a bug?

TF 版本:1.0.0

代码:

import tensorflow as tf

global_step = tf.Variable(0, dtype=tf.int64)
train_op = tf.assign(global_step, global_step + 1)
learning_rate = tf.Variable(0.1, dtype=tf.float32, name='learning_rate')

# Update the learning_rate tensor conditionally
# When global_step == 2, update to 0.01
# When global_step == 4, update to 0.001
cases = []
case_tensors = []
for step, new_rate in [(2, 0.01), (4, 0.001)]:
    pred = tf.equal(global_step, step)
    fn_tensor = tf.constant(new_rate, dtype=tf.float32)
    cases.append((pred, lambda: fn_tensor))
    case_tensors.append((pred, fn_tensor))
update = tf.case(cases, default=lambda: learning_rate)
updated_learning_rate = tf.assign(learning_rate, update)

print tf.__version__
with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())
    for _ in xrange(6):
        print sess.run([global_step, case_tensors, update, updated_learning_rate])
        sess.run(train_op)

结果:

1.0.0
[0, [(False, 0.0099999998), (False, 0.001)], 0.1, 0.1]
[1, [(False, 0.0099999998), (False, 0.001)], 0.1, 0.1]
[2, [(True, 0.0099999998), (False, 0.001)], 0.001, 0.001]
[3, [(False, 0.0099999998), (False, 0.001)], 0.001, 0.001]
[4, [(False, 0.0099999998), (True, 0.001)], 0.001, 0.001]
[5, [(False, 0.0099999998), (False, 0.001)], 0.001, 0.001]

推荐答案

This was answers in https://github.com/tensorflow/tensorflow/issues/8776

This was answered in https://github.com/tensorflow/tensorflow/issues/8776

事实证明，如果在 fn_tensors 中，lambda 返回一个在 lambda 之外创建的张量，则 tf.case 行为是未定义的.正确的用法是定义 lambdas，使它们返回一个新创建的张量.

It turns out that tf.case behavior is undefined if, in fn_tensors, the lambdas return a tensor that was created outside of the lambda. The correct usage is to define the lambdas such that they return a newly-created tensor.

根据链接的 Github 问题，这种用法是必需的，因为 tf.case 必须自己创建张量，以便将张量的输入连接到谓词的正确分支.

According to the linked Github issue, this usage is required because tf.case must create the tensor itself in order to hook up the tensor's inputs to the correct branch of the predicate.

这篇关于Tensorflow:为什么 tf.case 给我错误的结果?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！