tensorflow.python.framework.errors_impl.NotFoundError: ./backward_warp_op.so: undefined symbol: __cudaPushCallConfiguration

[MrYANG23]

Hi thanks your work ,
when i run run.py --help i got this error :tensorflow.python.framework.errors_impl.NotFoundError: ./backward_warp_op.so: undefined symbol: __cudaPushCallConfiguration
Traceback (most recent call last):
File "/home/yh/下载/UnFlow-master/src/e2eflow/ops.py", line 59, in
op_lib = tf.load_op_library(lib_path)
File "/home/yh/anaconda3/lib/python3.6/site-packages/tensorflow/python/framework/load_library.py", line 56, in load_op_library
lib_handle = py_tf.TF_LoadLibrary(library_filename)
tensorflow.python.framework.errors_impl.NotFoundError: ./backward_warp_op.so: cannot open shared object file: No such file or directory
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/yh/下载/UnFlow-master/src/run.py", line 7, in
from e2eflow.core.train import Trainer
File "/home/yh/下载/UnFlow-master/src/e2eflow/core/train.py", line 11, in
from . import util
File "/home/yh/下载/UnFlow-master/src/e2eflow/core/util.py", line 2, in
from ..ops import downsample as downsample_ops
File "/home/yh/下载/UnFlow-master/src/e2eflow/ops.py", line 62, in
op_lib = tf.load_op_library(lib_path)
File "/home/yh/anaconda3/lib/python3.6/site-packages/tensorflow/python/framework/load_library.py", line 56, in load_op_library
lib_handle = py_tf.TF_LoadLibrary(library_filename)
tensorflow.python.framework.errors_impl.NotFoundError: ./backward_warp_op.so: undefined symbol: __cudaPushCallConfiguration

can you give me some advices ,thank you sou much!

[AllenDun]

I run eval_gui.py, also got errors:
1 error detected in the compilation of tmp/tmpxft_00004b70_00000000-6_backward_warp_op.cu.cpp1.ii"
tensorflow.python.framework.errors_impl.NotFoundError: ./backward_warp_op.so: cannot open shared object file: No such file or directory
subprocess.CalledProcessError: Command 'nvcc -std=c++11 -c -o backward_warp_op.cu.o backward_warp_op.cu.cc -I/home/dfc/anaconda3/lib/python3.6/site-packages/tensorflow/include -D_GLIBCXX_USE_CXX11_ABI=0 -L/home/dfc/anaconda3/lib/python3.6/site-packages/tensorflow -ltensorflow_framework -D GOOGLE_CUDA=1 -x cu -Xcompiler -fPIC -I /usr/local --expt-relaxed-constexpr' returned non-zero exit status 1.

My OS is gcc (Ubuntu 5.4.0-6ubuntu1~16.04.10) 5.4.0 20160609 and tensorflow1.12. Is it related to gcc version or not? Thanks

[Fanpanda]

Hi, do you solve this problem? @AllenDun @MrYANG23

[MrYANG23]

Hi, do you solve this problem? @AllenDun @MrYANG23
i didnt solve this problem,i am trying reimplement this repo with pytorch again

[Fanpanda]

Thanks for your reply.

[3bhady]

@AllenDun I have the same issue. Have you had any luck with solving it?
@MrYANG23 did you implement it with pytorch?

[simonmeister]

Which version of cuda and cudnn are you using? After a quick search, it may be a problem with newer versions similar to pytorch/pytorch#15433.

[3bhady]

I'm running it on colab
Cuda compilation tools, release 10.0, V10.0.130

[aadilmehdis]

Was anyone able to fix this?

[simonmeister]

Did you try it with version 1.2 or 1.7 of tensorflow? I wasn't able to test a more recent version due to high workload from new projects.