yqyao / SSD_Pytorch

support different SSDs and different scale test, support refineDet.
MIT License
148 stars 51 forks source link

How to train in my own dataset? #20

Open haochange opened 5 years ago

haochange commented 5 years ago

I don‘t know where I should change the value? I changed: NUM_CLASSES in configuration.yaml VOC_CLASSES in voc0712.py

Are there others to be changed?

yqyao commented 5 years ago

you need change NUM_CLASSES in configuration.yaml @haochange

haochange commented 5 years ago

you need change NUM_CLASSES in configuration.yaml @haochange

Thank you for your timely reply @yqyao . I'll try and provide training logs for discussion.

yongqiang1210 commented 5 years ago

你好,能不能将具体的2012_test.txt, 2007_test.txt, 0712_trainval_test.txt, 2012_trainval.txt, 0712_trainval.txt的生成方式说一下。包括:那些文件生成对应的文件、怎么合并。——致谢

yqyao commented 5 years ago

readme 已经说了,就是把原始的VOC2007 VOC2012 ImageSet/Main/train.txt val.txt 等等 直接cat 合成一个大的 @ @yongqiang1210

yongqiang1210 commented 5 years ago

我按照您文件夹的意思进行了合并,但是VOC2012中没有_test.txt文件?如何合并生成2012_test.txt

yongqiang1210 commented 5 years ago

在运行代码时会显示: [Errno 2] No such file or directory: 'data/datasets/VOCdevkit0712/VOC0712/Annotations/002369 -1.xml'的错误,但是有002369.xml文件,没有后面的信息(1或者-1)。希望问一下是怎么回事?

yqyao commented 5 years ago

你在代码对应位置打印一下信息,看看是不是你的list写错了 @yongqiang1210

yongqiang1210 commented 5 years ago

好的

yongqiang1210 commented 5 years ago

您好,请问我用两个GPU去跑,GPU利用率忽高忽低,CPU占用率高达98%是为什么?能从哪里进行修改吗?——致谢

rw1995 commented 5 years ago

您好,请问我用两个GPU去跑,GPU利用率忽高忽低,CPU占用率高达98%是为什么?能从哪里进行修改吗?——致谢

能加个好友吗?我想尝试下如何调试这个,但是发现还是不对