是否需要从源头构建tensorflow才能使用初始？(Is it required to build tensorflow from source to use inception?)

是否需要从源头构建tensorflow才能使用初始？(Is it required to build tensorflow from source to use inception?)

我已经在他们的仓库中问过这个问题 ，但是他们在这里将堆栈溢出重定向到了我。 那么，是否需要从源头构建tensorflow才能使用？ 或者使用tensorflow的二进制安装足以使用tensorflow ？ 提前致谢。

I have already asked this question in their repository, but they redirected me here on stack overflow. So, is it required to build tensorflow from the sources to use inception? Or a binary install of tensorflow enough to use inception? Thanks in advance.

最满意答案

它从二进制安装正常工作。

tensorflow-models回购有一个示例笔记本，其中包含需要运行的样板代码。

在主站点的教程中也有一个例子 。

It works fine from the binary install.

The tensorflow-models repo has an example notebook with the boiler-plate code you need to get it running.

There is also an example in the tutorials on the main site.