1.运行fetch_data/fetch_caffe_mex_windows_vs2013_cuda65.m,无法下载,
官方百度链接下载解压放到external/caffe文件夹就行
结构为external/caffe/matlab
2.运行faster_rcnn_build.m报错
错误使用 mex
libcpmt.lib(xthrow.obj) : error LNK2038: 检测到“_MSC_VER”的不匹配项: 值“1900”不匹配值“1800”(nms_gpu_mex.o 中)
libcpmt.lib(syserror.obj) : error LNK2038: 检测到“_MSC_VER”的不匹配项: 值“1900”不匹配值“1800”(nms_gpu_mex.o 中)
正在创建库 D:\faster_rcnn-master\faster_rcnn-master\bin\nms_gpu_mex.lib 和对象
D:\faster_rcnn-master\faster_rcnn-master\bin\nms_gpu_mex.exp
nms_gpu_mex.o : error LNK2019: 无法解析的外部符号 "char const * __cdecl std::_Winerror_map(int)" (?_Winerror_map@std@@YAPEBDH@Z),该符号
在函数 "public: virtual class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > __cdecl
std::_System_error_category::message(int)const "
(?message@_System_error_category@std@@UEBA?AV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@2@H@Z) 中被引用
D:\faster_rcnn-master\faster_rcnn-master\bin\nms_gpu_mex.mexw64 : fatal error LNK1120: 1 个无法解析的外部命令
解决办法:找到
D:\faster_rcnn-master\faster_rcnn-master\functions\nms\nvmex.m
更改第8行
Host_Compiler_Location = '-ccbin "C:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\bin\x86_amd64"';
为自己vs的版本,比如系统要求1900,就要改为14
vs版本对应关系
Host_Compiler_Location = '-ccbin "C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\bin\x86_amd64"';
3.startup.m
fast_rcnn startup done
-------------Preparation for Testing:工作完成---------
接下来Testing Demo:
4.Run fetch_data/fetch_faster_rcnn_final_model.m to download our trained models.
不错意外又爆错,
fetch_faster_rcnn_final_model
Downloading faster_rcnn_final_model…
Error in downloading, please try links in README.md https://github.com/ShaoqingRen/faster_rcnn>>
直接百度下载再解压到相应文件目录
5.Run experiments/script_faster_rcnn_demo.m to test a single demo image.
fast_rcnn startup done
警告: 已为类 “double” 的值编制索引且未指定任何下标。当前此操作的结果是已编制索引的值本身,但在以后的版本中,它将成为错误。
In auto_select_gpu (line 16)
In script_faster_rcnn_demo (line 9)
GPU 1: free memory 3483368652
Use GPU 1
Invalid MEX-file ‘D:\faster_rcnn-master\faster_rcnn-master\external\caffe\matlab\caffe_faster_rcnn\caffe_.mexw64’: 找不到指定的模块。
可能caffe要重新编,直接下载的不能直接用 If not using VS 2013, you need to re-build the solution in .\Library\leveldb.
Then copy the built leveldb.lib to .\x64\Release_Mex.
If you are not using OpenCV 2.4.9, copy the corresponding opencv dll files to .\x64\Release_Mex.
caffe搭建
- 下载caffe-master,链接:http://pan.baidu.com/s/1qYKBEI4 密码:avlu
- 打开caffe-master文件夹,然后看到一个windows文件夹,然后继续打开windows文件夹,看到里面一个CommonSettings.props.example文件,复制出来一份,并改名字为CommonSettings.props。完成过后如下图:参考安装
- 修改 CommonSettings.props 文件,参考:https://blog.csdn.net/seu_nuaa_zc/article/details/80548998 主要是matlab、python、cuda、cudnn等等