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

关于多标签分类的问题 #13

Open
konioyxgq opened this issue Sep 4, 2019 · 6 comments
Open

关于多标签分类的问题 #13

konioyxgq opened this issue Sep 4, 2019 · 6 comments

Comments

@konioyxgq
Copy link

你好,我看你在多标签分类中用到了B-CNN,在VehicleDC.py中这样写的:
“X = torch.bmm(X, torch.transpose(X, 1, 2)) / (1 ** 2) # Bi-linear CNN”
我看了B-CNN的代码,并没有看到这样的方法。
请问你这样做的根据和目的是什么?
也可能是我看漏掉了,还请作者指点。非常感谢

@konioyxgq
Copy link
Author

@CaptainEven

@CaptainEven
Copy link
Owner

CaptainEven commented Sep 5, 2019

@konioyxgq 使用的是简化版的BCNN,两个分支共用一个网络,通过batch matrix multiply(torch.bmm)实现矩阵外乘。

@konioyxgq
Copy link
Author

@CaptainEven 谢谢你的回复,你的BCNN的文章链接好像错了,因该是http://vis-www.cs.umass.edu/bcnn/docs/bcnn_iccv15.pdf
而不是https://arxiv.org/pdf/1709.09890.pdf

@CaptainEven
Copy link
Owner

@konioyxgq 好的,多谢指正。

@konioyxgq
Copy link
Author

@CaptainEven 作者你好,我还想问一个关于BCNN的问题。我做了实验发现加入BCNN后和之前的结果比较,几乎差不多。不知道作者做过类似的试验吗?实验结果是什么样的呢?

@CaptainEven
Copy link
Owner

@konioyxgq BCNN主要用于细粒度分类,效果应该跟具体数据集有关,具体情况应该需要做更多的对比试验。

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