chandar-lab / recall2imagine Goto Github PK

When I use the default config.yaml to run the experiments of Memory Maze and POPGym, the following error occurs:
Process env: Traceback (most recent call last): File "/conda/miniconda/envs/py38/lib/python3.8/multiprocessing/process.py", line 315, in _bootstrap self.run() File "/conda/miniconda/envs/py38/lib/python3.8/multiprocessing/process.py", line 108, in run self._target(*self._args, **self._kwargs) File "/code/recall2imagine/embodied/core/distr.py", line 199, in _wrapper fn(*args) File "/code/recall2imagine/embodied/run/parallel.py", line 172, in env obs = env.step(act) File "/code/recall2imagine/embodied/core/wrappers.py", line 376, in step obs = self.env.step(action) File "/code/recall2imagine/embodied/core/wrappers.py", line 160, in step obs = self.env.step(action) File "/code/recall2imagine/embodied/core/wrappers.py", line 115, in step assert action[self._key].min() == 0.0, action AssertionError: {'action': array([nan, nan, nan, nan, nan, nan], dtype=float32), 'reset': array(False)}
It seems to be related to env reset. Could you give me some suggestions on why the error occurs and how to solve it? Looking forward to your reply.

I used the conda env and have installed jaxlib 0.4.13 with cuda11.cudnn86. When I run the code, the error occurred:
E external/xla/xla/stream_executor/cuda/cuda_dnn.cc:427] Loaded runtime CuDNN library: 8.5.0 but source was compiled with: 8.6.0. CuDNN library needs to have matching major version and equal or higher minor version. If using a binary install, upgrade your CuDNN library. If building from sources, make sure the library loaded at runtime is compatible with the version specified during compile configuration.