xbpeng / DeepMimic

Motion imitation with deep reinforcement learning.
https://xbpeng.github.io/projects/DeepMimic/index.html
MIT License
2.29k stars 484 forks source link

Can I work this in my mac #48

Open zdlarry opened 5 years ago

xbpeng commented 5 years ago

Sorry but we haven't tried building on macs

erwincoumans commented 5 years ago

Indeed, the rendering has some issues on Mac. I re-implemented the underlying environment from Bullet to PyBullet, which runs on Mac/Windows and Linux. It shares the same training code, arg files, mocap files and policies as DeepMimic.

pip3 install pybullet
python3 -m pybullet_envs.deep_mimic.testrl --arg_file run_humanoid3d_backflip_args.tx
zdlarry commented 5 years ago

@erwincoumans Thank you for your answer~. And can you tell me how to use pybullet to train with Multi-Clip Rewards such as target-heading and the training total time? Thank you very much!

nvcken commented 5 years ago
anaconda3/include/python3.7m/pystate.h:241:15: note: 'curexc_traceback' declared here
      PyObject *curexc_traceback;
                ^
  8 warnings and 15 errors generated.
  error: command 'gcc' failed with exit status 1

  ----------------------------------------
  Failed building wheel for PyOpenGL-accelerate
  Running setup.py clean for PyOpenGL-accelerate
  Running setup.py bdist_wheel for mpi4py ... error

I get this error when try to pip install on MacOS, any ideas

gaoalexander commented 3 years ago

Indeed, the rendering has some issues on Mac. I re-implemented the underlying environment from Bullet to PyBullet, which runs on Mac/Windows and Linux. It shares the same training code, arg files, mocap files and policies as DeepMimic.

pip3 install pybullet
python3 -m pybullet_envs.deep_mimic.testrl --arg_file run_humanoid3d_backflip_args.tx

Thank you for this, @erwincoumans !

I'm looking to diagnose the following error, wondering if you may have any idea what causes this:

OMP: Error #15: Initializing libiomp5.dylib, but found libomp.dylib already initialized.
OMP: Hint: This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade performance or cause incorrect results. The best thing to do is to ensure that only a single OpenMP runtime is linked into the process, e.g. by avoiding static linking of the OpenMP runtime in any library. As an unsafe, unsupported, undocumented workaround you can set the environment variable KMP_DUPLICATE_LIB_OK=TRUE to allow the program to continue to execute, but that may cause crashes or silently produce incorrect results. For more information, please see http://www.intel.com/software/products/support/.

Thank you very much.

EDIT: It seems that setting the Environment Variable KMP_DUPLICATE_LIB_OK=TRUE solves this issue.