search

YurongYou / Hindsight

Code release for "Hindsight is 20/20: Leveraging Past Traversals to Aid 3D Perception" [ICLR 2022]
MIT License
38 stars 4 forks source link

ValueError: cannot reshape array of size 265728 into shape (5) #2

Closed BiaoBiaoLi closed 2 years ago

BiaoBiaoLi commented 2 years ago

Thanks your work! when Download Lyft dataset and convert it to KITTI format, i meet this problem: File "/home/biaoli/anaconda3/envs/hindsight/lib/python3.8/site-packages/lyft_dataset_sdk/utils/data_classes.py", line 283, in from_file points = scan.reshape((-1, 5))[:, : cls.nbr_dims()] ValueError: cannot reshape array of size 265728 into shape (5) 24%|█████████▏ | 5471/22680 [03:31<11:04, 25.90it/s] is this a properly solution ? https://github.com/lyft/nuscenes-devkit/issues/89 thanks you !

YurongYou commented 2 years ago

Hi BiaoBiaoLi! Thanks for your interest! It seems I did not encounter this issue, but the solution you post should be proper (also see https://www.kaggle.com/c/3d-object-detection-for-autonomous-vehicles/discussion/110000).

YurongYou commented 2 years ago

I double-checked the data I used, and I found did patch the bin file before but forgot to include this step in the doc. Thanks for reporting it and I just updated the doc :)