Tangtao

3D boxes of instances in LIDAR coordinates about NuScenes

3

Hi, In nuSences the lidar coordinate is ( up z y front ^ ^ | / | / 0 ------> x), but the [gt_bbox_3d in nuscenes_dataset](https://github.com/open-mmlab/mmdetection3d/blob/eb5a5a2d166d2d60dfbdffe63b925cb37a6541e3/mmdet3d/datasets/nuscenes_dataset.py#L290) not change to the...

RuntimeError: cuda runtime error (2) when training

1

when I run train.py use my dataset, CUDA error: out of memory , how to solve it. My GPU is Titan X ,the batch is 2,and chunk size is [2]....

confuse about the depth supervision

2

Hi, thanks for your great work! In the code and tutorial, the depth supervision is conducted on the rendered_depth. But in paper the depth supervision is on ray termination, i.e.,...

Code for detection

4

Hi, thanks for your great work and the open source code ! If possible, I would greatly appreciate it if you could provide the code for 3D detection

Training process

1

Hi, sorry to bother you again. I'm confused how to integrate the computationally demanding reverse diffusion process, i.g, the step by step deniose process, into the perception train process. Is...

The difference with BEVFormer

1

Hi, Thanks for your open source work! But I have some doubts about the main differences between GKT and BEVFormer, except for the different selection of regions around the prior...

reimplementation problems on waymo

7

hello, thanks for your excellent work! But I have a problem with the reproduction of the waymo open dataset: I can get results of Transfusion-L: 'Overall/L1 mAP': 0.734978, 'Overall/L1 mAPH':...

hi, Thanks for your excellent work! However, when the number of points increased, I encountered a problem: ''' p1 = torch.rand(1,50000, 3).cuda() p2 = torch.rand(1,50000, 3).cuda() d = earth_mover_distance(p1, p2,...

Thank you for your great work and open source code. However, I am confused about the consistency loss part. 1.Could you please clarify how you calculate the pairwise distance between...