简体繁体 English

IBM Power9 ppc64le 上的 Tensorflow - 可以删除 libtensorflow.so 吗？

[英]Tensorflow on IBM Power9 ppc64le - Can libtensorflow.so be deleted?

原文 2020-10-21 14:31:59 0 1 python/ docker/ tensorflow/ conda/ powerpc

I tried to build a docker container with python and the tensorflow-gpu package for a ppc64le machine.我试图用 python 和 tensorflow-gpu 包为 ppc64le 机器构建一个 docker 容器。 I installed miniconda3 in the docker container and used the IBM repository to install all the necessary packages.我在 docker 容器中安装了 miniconda3 并使用 IBM 存储库安装了所有必要的包。 To my surprise the resulting docker container was twice as big (7GB) as its amd64 counterpart (3.8GB).令我惊讶的是，生成的 docker 容器是 amd64 对应容器（3.8GB）的两倍（7GB）。

I think the reason is, that the packages from the IBM repository are bloating the installation.我认为原因是 IBM 存储库中的软件包使安装变得臃肿。 I did some research and found two files libtensorflow.so and libtensorflow_cc.so in the tensorflow_core directory.我做了一些研究，在tensorflow_core目录中找到了两个文件libtensorflow.so和libtensorflow_cc.so 。 Both of theses files are about 900MB in size and they are not installed in the amd64 container.这两个文件的大小都约为 900MB，并且它们没有安装在 amd64 容器中。

It seems these two files are the API-files for programming with C and C++.这两个文件似乎是用于使用 C 和 C++ 编程的 API 文件。 So my question is: If I am planning on only using python in this container, can I just delete these two files or do they serve another purpose in the ppc64le installation of tensorflow?所以我的问题是：如果我打算只在这个容器中使用 python，我可以删除这两个文件还是它们在 tensorflow 的 ppc64le 安装中有其他用途？

1 个解决方案

Yes.是的。 Those are added as there were many requests for it and it's a pain to cobble together the libraries and headers yourself for an already built TF .whl.添加这些是因为有很多请求，并且为已经构建的 TF .whl 自己拼凑库和头文件是一种痛苦。

They can be removed if you'd rather have the disk space.如果您想要磁盘空间，可以将它们删除。

What is the content of your "amd64 container"?您的“amd64 容器”的内容是什么？ Just a pip install tensorflow?只是一个pip安装tensorflow？