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

source-only 训练,低mAP #47

Closed
M-o-magic opened this issue Jun 13, 2023 · 1 comment
Closed

source-only 训练,低mAP #47

M-o-magic opened this issue Jun 13, 2023 · 1 comment

Comments

@M-o-magic
Copy link

我在使用sigma代码时,想先用source-only预训练下试试看,所以把DA_ON关闭了。但是在source-only训练过程中,发现训练到40的mAP后,模型损失不再下降。我尝试过调整batch_size和learning_rate,但mAP就是上不去。当我使用yolov5进行sourceonly数据集的训练时,是可以到96的mAP的。
您对此问题有想法吗?

@wymanCV
Copy link
Contributor

wymanCV commented Jun 14, 2023

@M-o-magic 您好,谢谢对我们工作的关注.

您可以尝试下用官方FCOS代码试一下,其中trianing schedule与DA情况下不太一样.

另外请问您是使用自己的数据集吗?我们之前实验发现FCOS在某些其他数据集,比如Pascal VOC上的稳定性确实比Faster RCNN差一些,参数设置不好和其他检测器相比会出现比较大的性能差距.

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