chaytonmin / Occupancy-MAE

Official implementation of our TIV'23 paper: Occupancy-MAE: Self-supervised Pre-training Large-scale LiDAR Point Clouds with Masked Occupancy Autoencoders
Apache License 2.0
246 stars 18 forks source link

KeyError: 'road_plane' #26

Closed FrontierBreaker closed 1 year ago

FrontierBreaker commented 1 year ago

Is road_plane used during the pretraining for KITTI?

File "../pcdet/datasets/kitti/kitti_dataset.py", line 427, in getitem data_dict = self.prepare_data(data_dict=input_dict) File "../pcdet/datasets/dataset.py", line 130, in prepare_data 'gt_boxes_mask': gt_boxes_mask File "../pcdet/datasets/augmentor/data_augmentor.py", line 240, in forward data_dict = cur_augmentor(data_dict=data_dict) File "../pcdet/datasets/augmentor/database_sampler.py", line 252, in call data_dict = self.add_sampled_boxes_to_scene(data_dict, sampled_gt_boxes, total_valid_sampled_dict) File "../pcdet/datasets/augmentor/database_sampler.py", line 168, in add_sampled_boxes_to_scene sampled_gt_boxes, data_dict['road_plane'], data_dict['calib'] KeyError: 'road_plane'

chaytonmin commented 1 year ago

Is road_plane used during the pretraining for KITTI?

File "../pcdet/datasets/kitti/kitti_dataset.py", line 427, in getitem data_dict = self.prepare_data(data_dict=input_dict) File "../pcdet/datasets/dataset.py", line 130, in prepare_data 'gt_boxes_mask': gt_boxes_mask File "../pcdet/datasets/augmentor/data_augmentor.py", line 240, in forward data_dict = cur_augmentor(data_dict=data_dict) File "../pcdet/datasets/augmentor/database_sampler.py", line 252, in call data_dict = self.add_sampled_boxes_to_scene(data_dict, sampled_gt_boxes, total_valid_sampled_dict) File "../pcdet/datasets/augmentor/database_sampler.py", line 168, in add_sampled_boxes_to_scene sampled_gt_boxes, data_dict['road_plane'], data_dict['calib'] KeyError: 'road_plane'

No. Our voxel-MAE is self-supervised. No labels of gt boxes are used in training. You can refer to this error in OpenPCDet.

FrontierBreaker commented 1 year ago

I met this error when I run your code of voxel-mae pretraining on kitti. So you mean the road_plane is redundant to load in the code?

chaytonmin commented 1 year ago

I met this error when I run your code of voxel-mae pretraining on kitti. So you mean the road_plane is redundant to load in the code?

yes. We just keep the codes for simplicity.

FrontierBreaker commented 1 year ago

Got it. Thank you all the way!!