chengshuli / hrl4in Goto Github PK

iGibson hrl4in branch Installation failed.

While following the provided installation instruction, iGibson installation failed while installing pyopengl-accelerate.

I encountered some error when I'm trying to run the run_train_gibson.sh.

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/lu/anaconda3/envs/HRL4In/lib/python3.7/posixpath.py", line 94, in join
    genericpath._check_arg_types('join', a, *p)
  File "/home/lu/anaconda3/envs/HRL4In/lib/python3.7/genericpath.py", line 155, in _check_arg_types
    raise TypeError("Can't mix strings and bytes in path components") from None
TypeError: Can't mix strings and bytes in path components

I found that it was caused by the HRL4IN/iGibson/gibson2/core/render/mesh_renderer/mesh_renderer_cpu.py file in line 393-394.

dir = os.path.dirname(obj_path)
texture = loadTexture(os.path.join(dir, v), scale=texture_scale)

I found that the type of dir is string, while the type of v is bytes. And the command os.path.join(dir, v) caused the error.

The value of dir is "anaconda3/envs/HRL4In/lib/python3.7/site-packages/pybullet_data/roboschool/models_outdoor/stadium". I guessed it was a problem caused by the version of pybullet.

Hello authors, I used this ./run_train_gibson.sh script and I was not able to reproduce your results from the paper.

Can you please answer some questions regarding the model pretraining:

• Do you use a pre-trained low-level policy for training the HRL4IN algorithm? If yes can you please provide the weights of the learned policy?
• How do you train the only low-level policy?
  • What method (PPO, etc) do you use?
  • Do you train policy low-level policy separately for arm and base or combined?
    • If you train separately for arm and base then how do you combine those two learned policies?

Hi, Chengshu, I didn't find any Graphical display interface of Toy Env, is it because it wasn't made public？Thanks！

Thanks for your work.

I'm facing a problem when I run ./run_train.sh.

1. I intsalled GibsonEnvV2 successfully, and I run its examples/demo successfully.
2. I installed HRL4IN successfully.
   
3. However, when I run ./run_train.sh
   there is a error that recorded in .log file
   Traceback (most recent call last):
   File "train_hrl_gibson.py", line 13, in
   from gibson2learning.core.logging import logger
   ModuleNotFoundError: No module named 'gibson2learning'

My question is where is the "gibson2learning" package.
I can't find it in HRL4IN and GibsonEnvV2 project.

Can you give me some suggestions to solve this problem?
Thanks a lot!

Hello dear friends,

I am trying to install iGibson and ToyEnv according to the installation instructions. I applied all of instruction lines. Unfortunately, I dealed with lots of compatibility errors in conda env. Could you give exported file which contains versions of packages?

For example, I have a problem with torch version like following:

Thank you.