search

askforalfred / alfred

ALFRED - A Benchmark for Interpreting Grounded Instructions for Everyday Tasks
MIT License
375 stars 84 forks source link

RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED #23

Closed yuweijiang closed 4 years ago

yuweijiang commented 4 years ago

I met a problem with the cuDNN. My system is Unbuntu 18.04, my GPU is 2080Ti. I follow the requirement.txt to install the packages. Could you help me to solve this problem? Thanks a lot!

MohitShridhar commented 4 years ago

Hi @yuweijiang can you post the full error? Thanks!

yuweijiang commented 4 years ago

Sure, the full error is following.

{'tests_seen': 1533, 'tests_unseen': 1529, 'train': 21023, 'valid_seen': 820, 'valid_unseen': 821} Traceback (most recent call last): File "models/train/train_seq2seq.py", line 103, in model = model.to(torch.device('cuda')) File "/home/weijiang.yu/anaconda3/envs/alfred/lib/python3.6/site-packages/torch/nn/modules/module.py", line 386, in to return self._apply(convert) File "/home/weijiang.yu/anaconda3/envs/alfred/lib/python3.6/site-packages/torch/nn/modules/module.py", line 193, in _apply module._apply(fn) File "/home/weijiang.yu/anaconda3/envs/alfred/lib/python3.6/site-packages/torch/nn/modules/rnn.py", line 127, in _apply self.flatten_parameters() File "/home/weijiang.yu/anaconda3/envs/alfred/lib/python3.6/site-packages/torch/nn/modules/rnn.py", line 123, in flatten_parameters self.batch_first, bool(self.bidirectional)) RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED

MohitShridhar commented 4 years ago

@yuweijiang did you follow the installation instructions here?

Also, have you looked at this? You could try a smaller batch-size, as per the comments here.

Can you double check that these requirements with the right versions have been installed correctly in your virtualenv?