OpenPCDet
OpenPCDet copied to clipboard
OpenPCDet Toolbox for LiDAR-based 3D Object Detection.
@sshaoshuai Thanks for the awesome toolbox. But I'm not quite clear about why [this line](https://github.com/open-mmlab/OpenPCDet/blob/master/tools/train.py#L154) here occurs before the DDP model wrapper, even with the simple comment. Could you explain...
How to deal with the problem? When I using this "bash scripts/dist_train.sh ${NUM_GPUS} --cfg_file cfgs/waymo_models/centerpoint_4frames.yaml" to train waymo dataset. like this: multiprocessing/semaphore_tracker.py:144: UserWarning: semaphore_tracker: There appear to be 4 leaked...
Yes. I did inference the KITTI test set with official PV-RCNN weight, and submit the results to KITTI website. The results show as: Car (3D Detection) | 66.01 % |...
I saw results listed on table 16(PV-RCNN++) is different from original paper. Are the results reproduced here?
https://github.com/open-mmlab/OpenPCDet/blob/dd2edf9da0944238d3711f8145cc5b86c65f5a79/pcdet/ops/iou3d_nms/src/iou3d_nms_kernel.cu#L233 当A、B两个Box中存在负数边长时,其面积Sa和Sb可能被计算为负值。 因此此处的Sa+Sb-S_overlap结果可能为负值,并导致s_overlap / fmaxf(sa + sb - s_overlap, EPS) 为一个无意义的大数。 是否应该要确保sa、sb为正数,对其做abs处理?或与MMCV中的类似计算一样直接返回IoU=0? 参考MMCV的实现:https://github.com/open-mmlab/mmcv/blob/b9bdfeb6f6aa67372d18b2ed5a10de28fc179d11/mmcv/ops/csrc/common/box_iou_rotated_utils.hpp#L334
The evaluation results are weird as follows: 2022-09-10 13:24:55,922 INFO *************** Performance of EPOCH 30 ***************** 2022-09-10 13:24:55,923 INFO Generate label finished(sec_per_example: 0.0646 second). 2022-09-10 13:24:55,924 INFO recall_roi_0.3: 0.286138 2022-09-10...
I run the experiments on Openpcdet0.2, now I need to make the code correct on version 0.5. If I directly change the OpenPCDet/pcdet from 0.2 to 0.5 and re-setup, will...
Hi All, Thank you for this wonderful repo! We want to use OpenPCDet with ONCE dataset, do you plan to support it? If not do you have documentation / tutorial...
SST is an excellent 3D detection method based on transformer, but at present there is only implementation code based on mmdetection3d. (https://github.com/tusen-ai/SST) Can you provide SST support based on OpenPCDet?
Thanks the author share code about MPPNet. But based one the author readme, **First** I generated the new datasets about waymo, I want verify whether the dataset is true. This...