search

ChenYingpeng / caffe-yolov3

A real-time object detection framework of Yolov3/v4 based on caffe
472 stars 230 forks source link

大佬，我的电脑上没有cuda如何将cmake里面的cuda去掉啊 #78

Closed skybocai closed 3 years ago

skybocai commented 4 years ago

CUDA_CUDART_LIBRARY (ADVANCED) linked by target "yolov3-plugin" in directory /home/babi/Documents/caffe-yolov3 linked by target "demo" in directory /home/babi/Documents/caffe-yolov3/demo linked by target "eval" in directory /home/babi/Documents/caffe-yolov3/eval CUDA_TOOLKIT_INCLUDE (ADVANCED) used as include directory in directory /home/babi/Documents/caffe-yolov3 used as include directory in directory /home/babi/Documents/caffe-yolov3 used as include directory in directory /home/babi/Documents/caffe-yolov3

wycrystal commented 3 years ago

你好，请问你修改好了吗？我电脑上也没有cuda，如果你修改好了，可否告知一下，谢谢！

skybocai commented 3 years ago

放弃修改了，我直接改用了有cuda的电脑，如果你修改了可否让我也看一下怎么改的，谢谢

------------------ 原始邮件 ------------------ 发件人: "ChenYingpeng/caffe-yolov3" @.>; 发送时间: 2021年5月19日(星期三) 下午2:11 @.>; @.**@.>; 主题: Re: [ChenYingpeng/caffe-yolov3] 大佬，我的电脑上没有cuda如何将cmake里面的cuda去掉啊 (#78)

你好，请问你修改好了吗？我电脑上也没有cuda，如果你修改好了，可否告知一下，谢谢！

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

wycrystal commented 3 years ago

我也放弃了，直接又转换成其它模型测试了

alan8122 commented 3 years ago

很简单，就是把cuda的实现换成cup的实现，查看caffe的blob.cpp代码就会发现cuda使用的gpu_data() 会有一个cpu_data()对应，因此把YOLO实现中的gpu相关函数换成对应的cpu函数即可，按照这个思路就可以很快改出来

skybocai commented 3 years ago

谢谢大佬