Getting RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED while running `train_rl.sh`

zhejz / carla-roach

Roach: End-to-End Urban Driving by Imitating a Reinforcement Learning Coach. ICCV 2021.

https://zhejz.github.io/roach

Other

274 stars 50 forks source link

Getting RuntimeError: cuDNN error: CUDNN_STATUS_EXECUTION_FAILED while running `train_rl.sh` #21

Closed Z7MightGuy closed 1 year ago

Z7MightGuy commented 1 year ago

After I start running train_rl.sh, when the program executes this line self.policy.forward(self._last_obs) in ppo.py, the program gets stuck for a while and gives out the above error. @zhejz Do you have any idea?

zhejz commented 1 year ago

Hi, it should be related to cuDNN/cuda installation. Please double check your gpu is working and make sure you have installed the environment correctly.

atg93 commented 1 year ago

Hi, I encountered the same problem. Downloading

torch-1.12.0+cu113-cp38-cp38-linux_x86_64.whl torchvision-0.13.0+cu113-cp38-cp38-linux_x86_64.whl

in https://download.pytorch.org/whl/cu113/torch/ solved the problem in my case.