search

megvii-research / PETR

[ECCV2022] PETR: Position Embedding Transformation for Multi-View 3D Object Detection & [ICCV2023] PETRv2: A Unified Framework for 3D Perception from Multi-Camera Images
Other
879 stars 132 forks source link

i can't load mmdet3d_nuscenes_30f_infos_train.pkl #126

Open dsw-jlu-rgzn opened 1 year ago

yingfei1016 commented 1 year ago

Hi, The, mmdet3d_nuscenes_30f_infos_train.pkl is so large that it requires more memory. If you want to run the temporal model, I suggest running https://github.com/exiawsh/StreamPETR.git. It has lower requirements for the environment and higher accuracy. If you want to run segmentation, you can run with mmdet3d_nuscenes_2f_infos_train.pkl https://drive.google.com/drive/folders/1_C2yuh51ROF3UzId4L1itwGQVUeVUxU6.

dsw-jlu-rgzn commented 1 year ago

thanks~ And could you please provide a script on how to generate the temporal pkl files, I would like to refer to it.

dsw-jlu-rgzn commented 1 year ago

oh,I find the script generate_sweep_pkl.py^^

ysz-leo commented 5 months ago

The link to the script that generates the temporal pkl file(mmdet3d_nuscenes_30f_infos_train.pkl and mmdet3d_nuscenes_30f_infos_val.pkl) is at https://github.com/megvii-research/PETR/issues/4#issuecomment-1180171520