CornerNet-Lite top_pool.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN3c105ErrorC1ENS

top_pool.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN3c105ErrorC1ENS_14SourceLocationERKSs

Open lijunling opened this issue 5 years ago • 14 comments

master/core/models/py_utils/_cpools/init.py", line 6, in import top_pool, bottom_pool, left_pool, right_pool xxxx/lib/python3.7/site-packages/cpools-0.0.0-py3.7-linux-x86_64.egg/top_pool.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN3c105ErrorC1ENS_14SourceLocationERKSs my gcc is 6.1.0

May 06 '19 12:05 lijunling

Have you compiled sucessfully?

May 07 '19 02:05 lg920810

yes,when I run demo.py ,it throws the error.

May 07 '19 03:05 lijunling

you can try re-pull the code and build it again

May 07 '19 05:05 dmxj

I built the latest code, and also met question like this. ImportError: /root/.local/lib/python3.6/site-packages/cpools-0.0.0-py3.6-linux-x86_64.egg/top_pool.cpython-36m-x86_64-linux-gnu.so: undefined symbol: _ZTIN3c1021AutogradMetaInterfaceE

centos7 python 3.6.5 cuda 10.0 PyTorch 1.0.0 gcc 5.3.1

May 07 '19 06:05 arcral

I have same problem when run demo.py. Traceback (most recent call last): File "demo.py", line 4, in from core.detectors import CornerNet_Saccade File "/root/wcg/CornerNet-Lite-master/core/detectors.py", line 1, in from .base import Base, load_cfg, load_nnet File "/root/wcg/CornerNet-Lite-master/core/base.py", line 3, in from .nnet.py_factory import NetworkFactory File "/root/wcg/CornerNet-Lite-master/core/nnet/py_factory.py", line 7, in from ..models.py_utils.data_parallel import DataParallel File "/root/wcg/CornerNet-Lite-master/core/models/py_utils/init.py", line 1, in from ._cpools import TopPool, BottomPool, LeftPool, RightPool File "/root/wcg/CornerNet-Lite-master/core/models/py_utils/_cpools/init.py", line 6, in import top_pool, bottom_pool, left_pool, right_pool ImportError: /root/.local/lib/python3.7/site-packages/cpools-0.0.0-py3.7-linux-x86_64.egg/top_pool.cpython-37m-x86_64-linux-gnu.so: undefined symbol: _ZN2at19UndefinedTensorImpl10_singletonE

I change the pytorch to 0.4.1 because my cuda is 9.0.

May 09 '19 03:05 wcg5262

How to solve it? Thanks

May 13 '19 09:05 SeeeeShiwei

I have the same error. gcc 5.4.0

May 14 '19 05:05 andyhahaha

If comes with this error:undefined symbol: _ZN3c105ErrorC1ENS_14SourceLocationERKSs, follow 'Getting started' to create a new environment. Try not to run in your old environment, because it's easy to meet environment incompatible problem. I ran the demo firstly in my own environment and it appears error. Then I create a new environment and do as the 'Getting started' told, it works succesfully.