fastbev-mini v1.0

[bingxinhu]

nuscenes-mini

1.基本环境安装 conda create -n fastbev python=3.8 conda activate fastbev pip install torch==1.10.0+cu113 torchvision==0.11.0+cu113 torchaudio==0.10.0 -f https://download.pytorch.org/whl/torch_stable.html

安装mmcv-full安装时终端会卡住，不是卡住，是下载时间比较长，耐心等待

pip install mmcv-full==1.4.0

安装mmdet

pip install mmdet==2.14.0

安装mmdet

pip install mmsegmentation==0.14.1

2.下载fastbev工程 git clone https://github.com/Sense-GVT/Fast-BEV.git

激活虚拟环境

conda activate fastbev

conda install setuptools==58.0.4

进入Fast-BEV

cd Fast-BEV

配置所需依赖包

pip install -v -e .

or "python setup.py develop"

3.查看版本 python -c 'import mmcv;import mmdet;import mmdet3d;import mmseg;print(mmcv.version);print(mmdet.version);print(mmdet3d.version);print(mmseg.version)'

4.准备数据集 5.数据集转换为FastBEV支持格式 python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./data/nuscenes --extra-tag nuscenes --workers 10 --version v1.0-mini 生成索引 python tools/data_converter/nuscenes_seq_converter.py 分布式训练 python tools/train.py configs/fastbev/exp/paper/fastbev_m0_r18_s256x704_v200x200x4_c192_d2_f4.py --work-dir work_dir --gpu-ids 0 PC单卡训练 dist_train.sh configs/fastbev/exp/paper/fastbev_m0_r18_s256x704_v200x200x4_c192_d2_f4.py 1 推理测试 python tools/test.py configs/fastbev/exp/paper/fastbev_m0_r18_s256x704_v200x200x4_c192_d2_f4.py pretrained_models/epoch_20.pth --out output/result.pkl 可视化 python tools/misc/visualize_results.py configs/fastbev/exp/paper/fastbev_m0_r18_s256x704_v200x200x4_c192_d2_f4.py --result output/result.pkl --show-dir show_dir

具体详细操作参见 https://blog.csdn.net/h904798869/article/details/130317240