clvrai / goal_prox_il Goto Github PK

Hi! Thank you for the code.
I ran the provided code to check the experimental results on all scenarios except "Maze2D".
Fetch push, Navigation, Hand rotate scenarios showed same result with the paper.
But Fetch pick and Ant reach showed poor results, and appeared to not learn properly.

Here are the results.
I ran the exact same command from the readme.

Fetch pick

Also, I changed expert dataset to pick_partial.pt, it was better but still not showing the same result with the paper.

I don't think the virtual environment is wrong, because most scenarios had good results.
Maybe the code or the expert dataset is wrong with fetch pick and ant reach scenario 🤔

Hi,
Does the repo contain script to generate data for minigrid env? I wanted to understand how nav datasets were generated.

Hi!

Thank you for providing implementation! I am trying to use the repo, but I encoutered several errors when install mujoco_py and mpi4py. The error message is quite long, I selected some of them:

Building wheels for collected packages: mpi4py, mujoco-py
  Building wheel for mpi4py (setup.py) ... error
  error: subprocess-exited-with-error
  
  × python setup.py bdist_wheel did not run successfully.
  │ exit code: 1
  ╰─> [276 lines of output]

  note: This error originates from a subprocess, and is likely not a problem with pip.
  ERROR: Failed building wheel for mpi4py
  Running setup.py clean for mpi4py
  Building wheel for mujoco-py (setup.py) ... error
  error: subprocess-exited-with-error
  
  × python setup.py bdist_wheel did not run successfully.
  │ exit code: 1
  ╰─> [61 lines of output]
      running bdist_wheel
      running build

error: command '/usr/bin/mpicc' failed with exit code 1
[end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
error: legacy-install-failure

× Encountered error while trying to install package.
╰─> mpi4py

It seems to be caused by buliding process. Do you have any idea about the environment? Thx!