yukang

HI, Thanks for you information. I will finish adding it before the opening of CVPR 2022.

I have added this repo to Huggingface at here https://huggingface.co/Yukang/FocalsConv.

Hi, The original Voxel R-CNN receives point clouds only as input. Its config is https://github.com/open-mmlab/OpenPCDet/blob/master/tools/cfgs/kitti_models/voxel_rcnn_car.yaml In addition, we modify it to be multi-modal in this config, https://github.com/open-mmlab/OpenPCDet/blob/master/tools/cfgs/kitti_models/voxel_rcnn_car_focal_multimodal.yaml It receives both...

You can easily change other config files by `DATA_CONFIG` and `BACKBONE_3D` in [voxel_rcnn_car_focal_multimodal.yaml](https://github.com/open-mmlab/OpenPCDet/blob/master/tools/cfgs/kitti_models/voxel_rcnn_car_focal_multimodal.yaml) to make them supported by multi-modal inputs. For example, you can change https://github.com/open-mmlab/OpenPCDet/blob/master/tools/cfgs/kitti_models/pv_rcnn.yaml by the `DATA_CONFIG` and...

I met the same error.

你好，感觉不是很正常。我用4卡V100，BATCH_SIZE_PER_GPU=4. 训练大概2天多点。我猜测可能训练速度瓶颈在dataloader的数据读取上。

你好，我预估时间和实际时间基本差不多。

Hi, If you use only 1 GPU for testing, please use this script. `python3 test.py --cfg_file cfgs/nuscenes_models/cbgs_voxel0075_voxelnext.yaml --ckpt PATH_TO_VoxelNeXt.pth ` Regards, Yukang Chen

Have you installed OpenPCDet codebase? `python3 setup.py develop`

Hi, Would you please try to change `model_state_disk = checkpoint['model_state']` to `model_state_disk = checkpoint` in pcdet/models/detectors/detector3d_template.py? Regards, Yukang Chen