junjie18 / CMT

[ICCV 2023] Cross Modal Transformer: Towards Fast and Robust 3D Object Detection
Other
326 stars 35 forks source link

数据预处理报错ValueError: cannot reshape #110

Open Neko-coffee opened 1 month ago

Neko-coffee commented 1 month ago

在进行数据预处理的时候出现了以下报错,看样子是数组reshaped的问题,请问具体是什么原因导致的(没有使用提供的pkl文件) [>>>>>>>>>>>>>>>>>>> ] 22523/28130, 2.4 task/s, elapsed: 9411s, ETA: 2343sTraceback (most recent call last): File "tools/create_data.py", line 97, in nuscenes_data_prep( File "tools/create_data.py", line 47, in nuscenes_data_prep create_groundtruth_database(dataset_name, root_path, info_prefix, File "/data/coding/tools/data_converter/create_gt_database.py", line 105, in create_groundtruth_database example = dataset.pipeline(input_dict) File "/data/miniconda/envs/torch/lib/python3.8/site-packages/mmdet3d/datasets/pipelines/compose.py", line 49, in call data = t(data) File "/data/miniconda/envs/torch/lib/python3.8/site-packages/mmdet3d/datasets/pipelines/loading.py", line 228, in call points_sweep = np.copy(points_sweep).reshape(-1, self.load_dim) ValueError: cannot reshape array of size 229376 into shape (5)

Neko-coffee commented 1 month ago

最初使用过mini的数据集的时候没有出现过这个问题,这次用完整的数据集才出现了这个问题