BEVFusion(mit)源码复现时遇到的报错

最新推荐文章于 2025-03-20 18:34:18 发布

雨秒的對望

最新推荐文章于 2025-03-20 18:34:18 发布

阅读量1.9k

点赞数 16

文章标签：目标检测 python 人工智能深度学习

本文链接：https://blog.csdn.net/weixin_54678439/article/details/135061364

版权

总结在复现BEVFusion(mit)代码时训练、测试和可视化遇到的报错！

1、报错：ImportError: cannot import name ‘feature_decorator_ext’ from partially initialized module ‘mmdet3d.ops.feature_decorator’ (most likely due to a circular import) (/home/mengwen/bevfusion/mmdet3d/ops/feature_decorator/init.py)

解决：

（1）注释掉 mmdet3d/ops/init.py 中的 from .feature_decorator import feature_decorator

2、报错：FileNotFoundError: NuScenesDataset: [Errno 2] No such file or directory: ‘data/nuscenes//nuscenes_infos_train.pkl’

解决：更改tools/data_converter/nuscenes_converter.py 中第95~100行

info_path = osp.join(root_path,
'{}_infos_train.pkl'.format(info_prefix))
info_val_path = osp.join(root_path,
'{}_infos_val.pkl'.format(info_prefix))

3、报错：TypeError: FormatCode() got an unexpected keyword argument ‘verify’

解决：yapf包的版本问题，我原来yapf包的0.40.2，降低为yapf==0.40.1

pip install yapf==0.40.1

4、报错：AttributeError: module ‘distutils’ has no attribute ‘version’

解决：setuptools版本过高导致的问题

pip install setuptools==58.0.4

5、报错：RuntimeError: Given groups=1, weight of size [8, 1, 1, 1], expected input[24, 6, 256, 704] to have 1 channels, but got 6 channels instead

解决：将 mmdet3d/models/vtransforms/base 中第38行: add_depth_features=True 改为 False ,37行也改为 False

6、报错：RuntimeError: CUDA out of memory. Tried to allocate 168.00 MiB (GPU 0; 23.70 GiB total capacity; 2.41 GiB already allocated; 96.50 MiB free; 2.43 GiB reserved in total by PyTorch) If reserved memory i

解决：单卡gpu不能满足，采用分布式训练，在mmdet3d/apis/train.py文件中把distributed参数设置为True，直接vs-code全局搜索找到distributed

def train_model(
    model,
    dataset,
    cfg,
    distributed=True,
    validate=False,
    timestamp=None,
):

7、报错：ModuleNotFoundError: No module named ‘torchpack.utils.tqdm’

解决：把tools/visualize.py文件中from torchpack.utils.tqdm import tqdm改成from tqdm import tqdm