Tai-Wang
This seems to be caused by a subtle numerical problem. Because there can be negative values around zero in the results of `dist = a_square + b_square - 2 *...
Only need to fix minor comments. Others LGTM.
@filaPro Great contribution! BTW, I am curious about the iou computation used in mmdet3d. Sometimes I feel like there are also some corner cases that BEV IoU can not be...
> Do you have plans to release the code for your FCOS3D on NuScenes in near future? Yes. More information could be referred to #247 .
@lianqing11 Please share your idea here if you have any successful experiences afterward.
Seems not a bug related to spconv. It reports a representative error that is caused by not correctly compiling mmdet3d. So please check this more carefully.
Waiting for @VVsssssk final check
Hi @zzj403, thanks for your contribution! Could you please use pre-commit hook following this [guide](https://github.com/open-mmlab/mmcv/blob/master/CONTRIBUTING.md) to fix the linting errors?
We do not have more resources to support it soon. PRs from the community will be definitely welcome.
The automatically saving results of nuScenes evaluation should contain the PR curve figures. You can find them in the corresponding log directories.