Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

改了一大堆小问题,用voc数据集2个类型报类别错误 #12

Open
fenghuoxiguozu opened this issue Jul 7, 2022 · 2 comments

Comments

@fenghuoxiguozu
Copy link

image
f5a0c2838c2cd1b57c1dc8c9d2c6f38
loss回传时突然从tensor转成地址了

@gengyanlei
Copy link
Owner

gengyanlei commented Aug 8, 2022

@fenghuoxiguozu
sorry,才看到,一直没有登录。
这个仓库代码是没有经过测试的,所以可能会有很多问题。如果你基于voc训练,建议你从github上搜1个在voc上训练的star比较高,效果比较好的模型;或者去paper with code上查看在voc数据集上表现优异的模型及其代码链接。
我这里主要是想实现数据增强,但是由于工作原因,未完成测试,仅提供数据增强的一些idea。

@gengyanlei
Copy link
Owner

另外voc里面有255这种边界框,每个区域的边界都是这种,有专门处理voc的,就是根据[N,C,H,W]把边界框过滤掉在loss计算时变成[N,X]的形式计算,我这里没有考虑他的这种情况。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants