open-mmlab / mmsegmentation

OpenMMLab Semantic Segmentation Toolbox and Benchmark.
https://mmsegmentation.readthedocs.io/en/main/
Apache License 2.0
8.16k stars 2.6k forks source link

When I ran the code,I met some challenges #942

Closed 459737087 closed 3 years ago

459737087 commented 3 years ago

Thanks for your error report and we appreciate it a lot.

Checklist

I've searched all the issues,But I don't find the same problems

Describe the bug when I ran the code, It reported "ImportError: libcudart.so.10.1: cannot open shared object file: No such file or directory"

File "/usr/local/lib/python3.8/site-packages/mmseg/apis/init.py", line 2, in from .inference import inference_segmentor, init_segmentor, show_result_pyplot File "/usr/local/lib/python3.8/site-packages/mmseg/apis/inference.py", line 9, in from mmseg.models import build_segmentor File "/usr/local/lib/python3.8/site-packages/mmseg/models/init.py", line 2, in from .backbones import * # noqa: F401,F403 File "/usr/local/lib/python3.8/site-packages/mmseg/models/backbones/init.py", line 5, in from .fast_scnn import FastSCNN File "/usr/local/lib/python3.8/site-packages/mmseg/models/backbones/fast_scnn.py", line 7, in from mmseg.models.decode_heads.psp_head import PPM File "/usr/local/lib/python3.8/site-packages/mmseg/models/decode_heads/init.py", line 5, in from .cc_head import CCHead File "/usr/local/lib/python3.8/site-packages/mmseg/models/decode_heads/cc_head.py", line 8, in from mmcv.ops import CrissCrossAttention File "/usr/local/lib/python3.8/site-packages/mmcv/ops/init.py", line 2, in from .bbox import bbox_overlaps File "/usr/local/lib/python3.8/site-packages/mmcv/ops/bbox.py", line 4, in ext_module = ext_loader.load_ext('_ext', ['bbox_overlaps']) File "/usr/local/lib/python3.8/site-packages/mmcv/utils/ext_loader.py", line 13, in load_ext ext = importlib.import_module('mmcv.' + name) File "/usr/local/lib/python3.8/importlib/init.py", line 127, in import_module return _bootstrap._gcd_import(name[level:], package, level) ImportError: libcudart.so.10.1: cannot open shared object file: No such file or directory

mmcv-full 1.3.13 python 3.8 cuda 10.2

I found there is some changes in CriscrossAttention,But I am not sure

dataset camvid config pspnet command python -m torch.distributed.launch --nproc_per_node=1 tools/train.py configs/pspnet/pspnet_r50-d8_512x1024_40k_cityscapes.py --work-dir /output/mmsegmentation_up/ --launcher pytorch

MengzhangLI commented 3 years ago

Hi, your config file is about cityscapes dataset but your actual dataset is CamVid which is not official supported. I think it may raise some errors.

Best,

MengzhangLI commented 3 years ago

And if it is convenient you could upgrade mmseg and mmcv to latest version because we fix some bugs. Hoping to receive your feedback!

459737087 commented 3 years ago

Thank you ,I'll try. But when I degrade mmcv and mmseg version,this problem is solved. I ran successfully the code some months ago,the version is changed, and this problem come out. So, I want to learn about whether you changed the CrissCrossAttention code.

MengzhangLI commented 3 years ago

Thank you ,I'll try. But when I degrade mmcv and mmseg version,this problem is solved. I ran successfully the code some months ago,the version is changed, and this problem come out. So, I want to learn about whether you changed the CrissCrossAttention code.

Hope this link is useful to you: https://github.com/open-mmlab/mmcv/commit/2a3d2d4a357c08cc24f3755cd2ed3b9e43ff8445