RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED

[Roziallegro]

Hello, I am having trouble loading trying out test.py. I am currently running it in a docker environment: FROM nvidia/cuda:11.1.1-cudnn8-devel-ubuntu20.04 with python3.7 and torch==1.8.0, torchvision==0.9.0.

Traceback (most recent call last):
  File "test.py", line 260, in <module>
    gaze, gaze_bias, feat_map = model1(img_datas)  # (b * t, c, 16, 8)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)
  File "/dfgaze/l2cs1.py", line 55, in forward
    x = self.conv1(x)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py", line 889, in _call_impl
    result = self.forward(*input, **kwargs)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/conv.py", line 399, in forward
    return self._conv_forward(input, self.weight, self.bias)
  File "/usr/local/lib/python3.7/dist-packages/torch/nn/modules/conv.py", line 396, in _conv_forward
    self.padding, self.dilation, self.groups)
RuntimeError: cuDNN error: CUDNN_STATUS_NOT_INITIALIZED

Could it be compatibility issues?