TensorFlow 中 Variable 和 get_variable 的区别 [英] Difference between Variable and get_variable in TensorFlow

问题描述

据我所知，Variable是创建变量的默认操作，而get_variable主要用于权重共享.

As far as I know, Variable is the default operation for making a variable, and get_variable is mainly used for weight sharing.

一方面，有人建议在需要变量时使用 get_variable 代替原始的 Variable 操作.另一方面，我只在 TensorFlow 的官方文档和演示中看到了 get_variable 的任何使用.

On the one hand, there are some people suggesting using get_variable instead of the primitive Variable operation whenever you need a variable. On the other hand, I merely see any use of get_variable in TensorFlow's official documents and demos.

因此我想知道一些关于如何正确使用这两种机制的经验法则.是否有任何标准"原则?

Thus I want to know some rules of thumb on how to correctly use these two mechanisms. Are there any "standard" principles?

推荐答案

我建议始终使用 tf.get_variable(...) -- 这将使重构您的如果您需要随时共享变量，请编写代码，例如在多 GPU 设置中(请参阅多 GPU CIFAR 示例).它没有任何缺点.

I'd recommend to always use tf.get_variable(...) -- it will make it way easier to refactor your code if you need to share variables at any time, e.g. in a multi-gpu setting (see the multi-gpu CIFAR example). There is no downside to it.

纯 tf.Variable 是低级的；在某些时候 tf.get_variable() 不存在所以一些代码仍然使用低级方式.

Pure tf.Variable is lower-level; at some point tf.get_variable() did not exist so some code still uses the low-level way.

这篇关于TensorFlow 中 Variable 和 get_variable 的区别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！