[Bug]: constant error about cudnn not supported

[yggdrasil75]

What happened?

gits/OneTrainer/venv/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: Plan failed with a cudnnException: CUDNN_BACKEND_EXECUTION_PLAN_DESCRIPTOR: cudnnFinalize Descriptor Failed cudnn_status: CUDNN_STATUS_NOT_SUPPORTED (Triggered internally at ../aten/src/ATen/native/cudnn/Conv_v8.cpp:919.) return Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass

every few steps. I am on linux mint 21 using kernel 6.5 with nvidia 545 and cuda 1.12.5. also python 3.10

What did you expect would happen?

train using cuda 1.12.5

Relevant log output

gits/OneTrainer/venv/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: Plan failed with a cudnnException: CUDNN_BACKEND_EXECUTION_PLAN_DESCRIPTOR: cudnnFinalize Descriptor Failed cudnn_status: CUDNN_STATUS_NOT_SUPPORTED (Triggered internally at ../aten/src/ATen/native/cudnn/Conv_v8.cpp:919.)
  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass

Output of pip freeze

freeze.txt