search

Toni-SM / skrl

Modular reinforcement learning library (on PyTorch and JAX) with support for NVIDIA Isaac Gym, Omniverse Isaac Gym and Isaac Lab
https://skrl.readthedocs.io/
MIT License
443 stars 43 forks source link

Getting blank stderr while training with OIGE #133

Closed Pipe-Runner closed 7 months ago

Pipe-Runner commented 8 months ago

Description

I get duplicate progress bars from stderr. This code used to work perfectly fine in the previous version of OIGE. I have upgraded to 2023.1.0 yesterday. I have been getting this behavior since then.

image The training seems to work correctly though. No error and the policy is trained as expected.

I do get this before the training starts though. Not sure if this is related to the problem. image

What skrl version are you using?

1.0.0

What ML framework/library version are you using?

PyTorch Version: 2.0.1, OIGE 2023.1.0a0

Additional system information

Linux (Ubuntu 22.04)

Toni-SM commented 8 months ago

Hi @Pipe-Runner

Thanks for reporting this issue. I have changed (in the develop branch, to be released soon) the output of the progress messages (generated with tdqm) to sys.stdout (tqdm default sys.stderr) so Isaac Sim will not show duplicated message anymore.

https://github.com/Toni-SM/skrl/blob/7b090a1674b3ff1240a49385bd7fd757571f7145/skrl/trainers/torch/base.py#L165

https://github.com/Toni-SM/skrl/blob/bb493a5c8e9ad71c0412d46df38802038ede4c86/skrl/trainers/torch/base.py#L166