Python Tensorflow：.ckpt 文件和 .ckpt.meta 和 .ckpt.index 和 .pb 文件的关系是什么

Tensorflow : What is the relationship between .ckpt file and .ckpt.meta and .ckpt.index , and .pb file

提问by Frank.Fan

I used saver=tf.train.Saver()to save the model that I trained, and I get three kinds of files named:

我曾经saver=tf.train.Saver()保存我训练过的模型，我得到了三种文件，命名为：

• .ckpt.meta
• .ckpt.index
• .ckpt.data

• .ckpt.meta
• .ckpt.index
• .ckpt.data

And a file called:

还有一个名为：

• checkpoint

• 检查站

What is the connection with the .ckptfile?

与.ckpt文件有什么联系？

I saw someone saved model with only .ckpt file, I don't know how to make it. How can I save model as a .pb file?

我看到有人保存的模型只有 .ckpt 文件，我不知道如何制作。如何将模型保存为 .pb 文件？

回答by gdelab

• the .ckpt file is the old version output of saver.save(sess), which is the equivalent of your .ckpt-data(see below)

• the "checkpoint" file is only here to tell some TF functions which is the latest checkpoint file.

• .ckpt-metacontains the metagraph, i.e. the structure of your computation graph, without the values of the variables (basically what you can see in tensorboard/graph).

• .ckpt-datacontains the values for all the variables, without the structure. To restore a model in python, you'll usually use the meta and data files with (but you can also use the .pbfile):

  saver = tf.train.import_meta_graph(path_to_ckpt_meta)
  saver.restore(sess, path_to_ckpt_data)
  

• I don't know exactly for .ckpt-index, I guess it's some kind of index needed internally to map the two previous files correctly. Anyway it's not really necessary usually, you can restore a model with only .ckpt-metaand .ckpt-data.

• the .pbfile can save your whole graph (meta + data). To load and use (but not train) a graph in c++ you'll usually use it, created with freeze_graph, which creates the .pbfile from the meta and data. Be careful, (at least in previous TF versions and for some people) the py function provided by freeze_graphdid not work properly, so you'd have to use the script version. Tensorflow also provides a tf.train.Saver.to_proto()method, but I don't know what it does exactly.

• .ckpt 文件是 的旧版本输出saver.save(sess)，相当于您的.ckpt-data（见下文）

• “checkpoint”文件只是在这里告诉一些TF函数哪个是最新的检查点文件。

• .ckpt-meta包含元图，即计算图的结构，没有变量的值（基本上是你可以在张量板/图中看到的）。

• .ckpt-data包含所有变量的值，没有结构。要在 python 中恢复模型，您通常会使用元和数据文件（但您也可以使用该.pb文件）：

  saver = tf.train.import_meta_graph(path_to_ckpt_meta)
  saver.restore(sess, path_to_ckpt_data)
  

• 我不确切知道 for .ckpt-index，我猜这是内部需要某种索引才能正确映射前两个文件。无论如何，通常并不是真的有必要，您可以仅使用.ckpt-meta和来恢复模型.ckpt-data。

• 该.pb文件可以保存您的整个图形（元 + 数据）。要在 C++ 中加载和使用（但不训练）图形，您通常会使用它，用 创build freeze_graph，它.pb从元和数据创建文件。请注意，（至少在以前的 TF 版本和某些人中）提供的 py 功能freeze_graph无法正常工作，因此您必须使用脚本版本。Tensorflow 也提供了一种tf.train.Saver.to_proto()方法，但我不知道它到底是做什么的。

There are a lot of questions here about how to save and restore a graph. See the answer herefor instance, but be careful that the two cited tutorials, though really helpful, are far from perfect, and a lot of people still seem to struggle to import a model in c++.

这里有很多关于如何保存和恢复图形的问题。例如，请参阅此处的答案，但请注意，这两个引用的教程虽然确实很有帮助，但远非完美，而且很多人似乎仍然难以在 C++ 中导入模型。

EDIT: it looks like you can also use the .ckpt files in c++ now,so I guess you don't necessarily need the .pb file any more.

编辑：看起来您现在也可以在 c++ 中使用 .ckpt 文件，所以我想您不一定需要 .pb 文件了。