简体   繁体   English

使用Windows10构建label_image张量流示例

[英]Build label_image tensorflow example with Windows10

I built tensorflow following every bit of instruction here . 我按照这里的每条指令构建了tensorflow。
The build with VS2015 x64 looks fine, but when compiling the tensorflow\\examples\\label_image sample I have many errors: VS2015 x64的构建看起来不错,但是在编译tensorflow\\examples\\label_image示例时,我有很多错误:

\tensorflow\third_party\eigen3\unsupported\eigen\cxx11\tensor(1): fatal error C1014: too many include files: depth = 1024

Fixing it with a #pragma once , then I got that this (and many like this) a #pragma oncea #pragma once修复它,然后我得到了这个(还有很多这样的东西)

template <>
struct NumTraits<QInt8> : GenericNumTraits<int8_t> {};

gives this error: 给出此错误:

\\tensorflow\\third_party\\eigen3\\unsupported\\eigen\\cxx11\\src\\fixedpoint\\fixedpointtypes.h(27): error C2988: unrecognizable template declaration/definition

I also miss the file 我也错过了文件

#include "tensorflow/cc/ops/image_ops.h"

Did anyone succeed to built (and link...) this sample? 有没有人成功构建(并链接...)此示例?

may be your include path is wrong. 可能是您的包含路径错误。 in my project i copy all incloud path in _beam_search_ops project to my project. 在我的项目中,我将_beam_search_ops项目中的所有incloud路径复制到我的项目中。 then the problem is solved 那么问题就解决了

I had the same issue, but this article cleared it up: 我遇到了同样的问题,但是这篇文章解决了这个问题:

https://joe-antognini.github.io/machine-learning/windows-tf-project https://joe-antognini.github.io/machine-learning/windows-tf-project

Your include directories need to be in the correct order, specifically 您的包含目录需要正确顺序,特别是

{your path}\\tensorflow\\tensorflow\\contrib\\cmake\\build\\external\\eigen_archive {您的路径} \\ tensorflow \\ tensorflow \\ contrib \\ cmake \\ build \\ external \\ eigen_archive

must be listed before 必须在之前列出

{your path}\\tensorflow\\third_party\\eigen3 {您的路径} \\ tensorflow \\ third_party \\ eigen3

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM