When I run test.py, cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

[qqpann]

Docker cuda10.0 ubuntu18.04
Python 3.7.14
nvcc -V -> 10.0
nvidia-smi -> 11.4

The full error:

root@db46231c9f61:/app/DCN# python test.py
THCudaCheck FAIL file=/pytorch/aten/src/THC/THCGeneral.cpp line=405 error=8 : invalid device function
Traceback (most recent call last):
  File "test.py", line 605, in <module>
    example_dconv()
  File "test.py", line 512, in example_dconv
    output = dcn(input)
  File "/root/.pyenv/versions/3.7.14/lib/python3.7/site-packages/torch/nn/modules/module.py", line 489, in __call__
    result = self.forward(*input, **kwargs)
  File "/app/DCN/modules/deform_conv.py", line 89, in forward
    offset = self.conv_offset(input)
  File "/root/.pyenv/versions/3.7.14/lib/python3.7/site-packages/torch/nn/modules/module.py", line 489, in __call__
    result = self.forward(*input, **kwargs)
  File "/root/.pyenv/versions/3.7.14/lib/python3.7/site-packages/torch/nn/modules/conv.py", line 320, in forward
    self.padding, self.dilation, self.groups)
RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

[hardbone6]

I have a same problem，Have you solved it？