StreamPETR

StreamPETR copied to clipboard

Inquiry on Incorporating Future Data in Training for StreamPETR.

6

Hello, I would like to seek your guidance on whether StreamPETR can also incorporate future data into training, similar to Sparse4D-v3-offline. If possible, could you kindly advise how I should...

kkkcx

[问题] gt_bboxes_3d是如何由sample_annotation.json转换得到的？

2

我尝试去理解目标框的ground truth，于是把data loader加载的一个sample的gt_bboxes_3d保存了下来，然后画到相应的相机图片上。 保存gt_bboxes_3d的位置： [petr3d.py#L223](https://github.com/exiawsh/StreamPETR/blob/main/projects/mmdet3d_plugin/models/detectors/petr3d.py#L223) 画图的方法 （我理解gt_bboxes_3d是lidar boxes 3d，是在sensor坐标系的，所以画到相机视角是要校准）： ``` box_list = [] for box in boxes_mine: # Move from censor coordinate to world coordinate box.rotate(Quaternion(lidar_rotation)) box.translate(np.array(lidar_translation)) box.rotate(Quaternion(pose_record['rotation'])) box.translate(np.array(pose_record['translation'])) if...

LiuJieShane

KeyError: 'Petr3D_seg: "PETRHead_seg: \'SinePositionalEncoding3D is not in the position encoding registry\'"'

3

When i try to add the segmentation task from PETRv2 to StreamPETR, i put【petrv2_ BEVseg.py, petr3d_Seg. py, petr_head_seg.py】 in the corresponding location under the StreamPETR project, i also add 【from...

Amireux52

about customdataset

6

hello, i want to use my dataset to train the streampetr. what do u think about that, 500 frames is enough to let the transformer convergent or not. the performance...

XLiedInfi

used data

8

Thanks for the grate work. i have aquestion regarding the used data: in the nuscenes2d_temporal_infos_train.pkl file there are many keys ['lidar_path', 'token', 'prev', 'next', 'sweeps', 'frame_idx', 'cams', 'scene_token', 'lidar2ego_translation', 'lidar2ego_rotation',...

RashoAli

How to convert streampetr pytorch model to onnx model?

1

wjmcidi

Can you provide onnx and deployment environment docker？

3

qingwan7

关于 flash attention版本导致 VOVNet 为backbone的收敛问题

2

想请问一下，当flash attention版本从 0.2 切换到 2.0, stream_petr_vov_flash_800_bs2_seq_24e (Vovnet为backbone) 收敛会有问题，loss和grad norm会出现 NAN的情况，导致这个的原因是什么呢，谢谢。

nancy618618

Running tracking evaluation

1

I want to run tracking evaluation using `nusc_tracking/pub_test`. However, I'm not sure what is the path I should provide to `--checkpoint`. Can you point me to the correct setup?

ekosman

Validity of using gray images :)

1

We want to use gray scale images to do inflection. Is this valid? Tell me what you think! Thank you!! @exiawsh

sjminInSwm