GitLab

* Add basic support for Pandaset dataset

* Add gt_sampling support for pandaset

* Enable training on different classes than raw dataset

* Save pandaset output in the same format as annotations

* Add dataset config file for pandaset

* Make pandaset dataset compatible with test.py

- add 'evaluation' method
- add 'name' key to predictions dict

* Enable selecting lidar device for pandaset

Co-authored-by: xiazhongyu <xiazhongyu@pku.edu.cn>

support FILTER_EMPTY_BOXES_FOR_TRAIN and DISABLE_NLZ_FLAG_ON_POINTS for WOD by default, eval with sample_interval=1

update create_groundtruth_database of WOD to generate global npy file for sharedmemory based GT_database

Improve WOD performance with two LiDAR returns, speed up WOD data pre-processing with multiprocesssing.Pool

bugfixed: weights transformation between spconv 1.x and spconv 2.x, should consider both native/implicit spconv 2.x

- as per the spconv documentation, the order of channels in weights is now changed
- we need to get all the affected weights, using the util function in spconv_utils
- we need to check the version with the version checker in __init__.py
- we need to transpose last two dimensions of existing weights, from older models
- to avoid duplication a single _load_state_dict method is added

missing kornia. before image_to_tensor

Improve WOD performance with two LiDAR returns, speed up WOD data pre-processing with multiprocesssing.Pool

* add voxel roi pooling

* add voxel r-cnn head

* add voxel query

* add voxel r-cnn

* add Voxel R-CNN

* add infos about Voxel R-CNN

* add voxel_rcnn_car.yaml

* add infos about Voxel R-CNN
Co-authored-by: Shaoshuai Shi <shaoshuaics@gmail.com>