We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
发现当label中大小目标差异比较大的时候,用f = iou > iou.mean()过滤后,很多小目标的label完全过滤掉,无法参加训练。比如我的label分别为0,1. 0是大的label,1是小的label。过滤后,只有0的label参加训练。1对应的被忽略
The text was updated successfully, but these errors were encountered:
请问解决了吗这个问题
Sorry, something went wrong.
+1
解决了嘛
No branches or pull requests
发现当label中大小目标差异比较大的时候,用f = iou > iou.mean()过滤后,很多小目标的label完全过滤掉,无法参加训练。比如我的label分别为0,1. 0是大的label,1是小的label。过滤后,只有0的label参加训练。1对应的被忽略
The text was updated successfully, but these errors were encountered: