mmdetection3d环境部署

安装python库

pip install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html

pip install mmengine
pip install mmcv-2.0.0rc4-cp38-cp38-manylinux1_x86_64.whl
pip install 'mmdet>=3.0.0,<3.1.0'
cd mmdetection3d
pip install -v -e .

python库检查

python -c 'import torch;print(torch.__version__);print(torch.version.cuda)'
1.9.1+cu111
11.1
python -c 'import mmengine;print(mmengine.__version__);'
0.10.6
python -c 'import mmcv;print(mmcv.__version__);'
2.0.0rc4
python -c 'import mmdet;print(mmdet.__version__);'
3.3.0
python -c 'import numpy;print(numpy.__version__);'
1.24.4

推理验证

下载模型权重

https://github.com/open-mmlab/mmdetection3d/tree/main/configs/pointpillars

模型推理

python demo/pcd_demo.py demo/data/kitti/000008.bin configs/pointpillars/pointpillars_hv_secfpn_8xb6-160e_kitti-3d-car.py models/hv_pointpillars_secfpn_6x8_160e_kitti-3d-car_20220331_134606-d42d15ed.pth --show

python demo/pcd_demo.py demo/data/nuscenes/n015-2018-07-24-11-22-45+0800__LIDAR_TOP__1532402927647951.pcd.bin* configs/pointpillars/pointpillars_hv_secfpn_sbn-all_8xb2-amp-2x_nus-3d.py models/hv_pointpillars_secfpn_sbn-all_fp16_2x8_2x_nus-3d_20201020_222626-c3f0483e.pth --show

python demo/pcd_demo.py demo/data/nuscenes/n015-2018-07-24-11-22-45+0800__LIDAR_TOP__1532402927647951.pcd.bin* configs/pointpillars/pointpillars_hv_secfpn_sbn-all_8xb4-2x_nus-3d.py models/hv_pointpillars_secfpn_sbn-all_fp16_2x8_2x_nus-3d_20201020_222626-c3f0483e.pth --show