Tensorflow Warning: The TensorFlow library wasn’t compiled to use SSE,SSE2,SSE3,SSE4.1 instructions,but these are available on your machine and could speed up CPU computations.

在使用tensorflow时，会一段巨长的warning，如下图所示：

总结来说，就是电脑上的tensorflow不能编译SSE, SSE2, SSE3, SSE4.1这些指令集, 但是机器上是有这些指令集的，并且使用这些指令会加快CPU的运行速度.

然而当前的TensorFlow在编译时并没有用到这些指令集，需要手动编译才能加入这些指令集. 所以tensorflow的运行速度会比较慢，不过程序仍然是可以运行的.

原因

出现这种warning的原因是当前使用的tensorflow是采用pip命令安装的，不是从源码编译安装的. 安装的Tensorflow中缺少了对应的模块，需要编译安装。
顺便提一句，当我们使用GPU时并不需要用到CPU，所以这些warning可以忽略.

解决

查阅网上的资料，发现要想消除这个warning有2种方法：

在代码前加上以下指令，修改警告信息的显示级别，使得运行不出现warning提示：

import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='2'
import tensorflow as tf

为什么加了这句指令就不会出现Warning提示了呢？因为修改了信息的显示级别：

# 1級
import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='1'    # 所有信息都显示
# 2級
import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='2'     # 只显示 warning 和 Error
# 3級
import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='3'     # 只显示 Error

重新编译安装tensorflow，在编译的时候使用这些指令集.
具体的步骤在其他博客中有写到，这里只规整一下：

Step1 卸载已经安装的tensorflow
Step2 克隆tensorflow仓库，使用下面的命令会在当前文件夹中创建一个名为“tensorflow”的文件夹，下载的文件都存在里面.

git clone --recurse-submodules https://github.com/tensorflow/tensorflow

Step3 安装bazel，在编译安装tensorflow的时候要用到bazel工具. 按照官网指导输入命令.
Step4 开始编译TensorFlow
a. configure

cd tensorflow
./configure

b. bazel build
#CPU

bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package

c.生成whl文件

bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg

d. 安装编译好的包

pip install /tmp/tensorflow_pkg/tensorflow-{version}-none-any.whl

e. 验证TensorFlow是否安装成功

>>>import tensorflow as tf
>>>hello = tf.constant("Hello!TensorFlow!")
>>>sess = tf.Session()
>>>print(sess.run(hello))

测试成功即可，我还没有试验是否可行，先记下来，找合适时间实践.

具体的过程在参考资料[1]中有着详细的讲解.
参考资料：
[1] TensorFlow CPU环境 SSE/AVX/FMA 指令集编译
sess.run()出现如下Warning

[2] TensorFlow問題“The TensorFlow library wasn’t compiled to use SSE instructions, but these are available on your machine and could speed up CPU conputations

Warning: The TensorFlow library wasn't compiled to use SSE,SSE2,SSE3,SSE4.1 instructions相关推荐

The TensorFlow library wasn‘t compiled to use SSE instructions, but these are available on your mach
pycharm python中遇到The TensorFlow library wasn't compiled to use SSE instructions, but these are avail ...
运行TensorFlow出现The TensorFlow library wasn‘t compiled to use FMA instructions）代码
TensorFlow出错点(2)(The TensorFlow library wasn't compiled to use FMA instructions) 错误提示: 解决措施: 错误提示: 摘 ...
Tensorflow警告：our CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
使用TensorFlow模块时,弹出错误Your CPU supports instructions that this TensorFlow binary was not compiled to u ...
去除警告：Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
去除警告:Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA 文章 ...
成功解决Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX AVX2(二)
成功解决Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX AVX2(二) ...
成功解决Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX AVX2
成功解决Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX AVX2 目录 ...
彻底解决“Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA”警告
问题描述在使用TensorFlow时,总是提醒"Your CPU supports instructions that this TensorFlow binary was not com ...
tensorflow:Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
ps:这跳出的是一个cpu警告,但我用的是gpu啊!!即使我安装合适版本的(编译过的)tensorflow来满足我cpu的AVX2扩展,也还是没对我的运行速度有太大提高(猜测),所以如果你是使用gpu ...
警告：Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA(亲测)
警告:Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA 转载:h ...

Warning: The TensorFlow library wasn't compiled to use SSE,SSE2,SSE3,SSE4.1 instructions

Tensorflow Warning: The TensorFlow library wasn’t compiled to use SSE,SSE2,SSE3,SSE4.1 instructions,but these are available on your machine and could speed up CPU computations.

原因

解决

Warning: The TensorFlow library wasn't compiled to use SSE,SSE2,SSE3,SSE4.1 instructions相关推荐

最新文章

热门文章