Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR in RNNLogic+ #17

Open
yzhangee opened this issue Jul 19, 2022 · 0 comments
Open

Comments

@yzhangee
Copy link

Dear Authors,

I tried to run this code on FB15k-237 on a single GPU (after revising config file) but face the following error. The related code for rnn seems correct. I searched this error online and did not fine a solution. So I wonder why this problem happens and how to solve it. My GPU is 3090 with CUDA 11.2.

Traceback (most recent call last):
  File "/home/Logic/RNNLogic/RNNLogic+/src/run_rnnlogic.py", line 143, in <module>
    main(parse_args())
  File "/home/Logic/RNNLogic/RNNLogic+/src/run_rnnlogic.py", line 58, in main
    solver_g.train(dataset, **cfg.generator.pre_train)
  File "/home/Logic/RNNLogic/RNNLogic+/src/trainer.py", line 327, in train
    loss = model.loss(inputs, target, mask, weight, hidden)
  File "/home/Logic/RNNLogic/RNNLogic+/src/generators.py", line 32, in loss
    logits, hidden = self.forward(inputs, inputs[:, 0], hidden)
  File "/home/Logic/RNNLogic/RNNLogic+/src/generators.py", line 27, in forward
    outputs, hidden = self.rnn(embedding, hidden)
  File "/home/anaconda3/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/anaconda3/lib/python3.9/site-packages/torch/nn/modules/rnn.py", line 691, in forward
    result = _VF.lstm(input, hx, self._flat_weights, self.bias, self.num_layers,
RuntimeError: cuDNN error: CUDNN_STATUS_INTERNAL_ERROR
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant