I need help on how to run Learned Policy in Isaac Lab direct workflow

Omniverse Isaac Sim
1

Dear All,

I need help, I have created a Low-Level Policy using Isaac Lab, but I am stuck with the RL code for a High-Level Policy using this policy.
I tried to reference the navigation demo code in the Manager Base Workflow, but I gave up because I am adopting the direct workflow. So I am working on diverting the play.py from rsl-rl, but am having difficulty.
Does anyone know how to run a learned Policy in Isaac Lab’s direct workflow?

Thank you!

Isaac Sim Version

4.2.0
4.1.0
4.0.0
2023.1.1
2023.1.0-hotfix.1
Other (please specify):

Isaac Lab Version (if applicable)

1.2
1.1
1.0
Other (please specify):

Operating System

Ubuntu 22.04
Ubuntu 20.04
Windows 11
Windows 10
Other (please specify):

GPU Information

  • Model: RTX A4000
  • Driver Version:
2

I’m just getting started in lab but hopefully the little I know can help here

in

.IsaacLab\source\standalone\workflows\rsl_rl

Run your training file it will produce the logs folder

image
image1179×332 13.6 KB

inside that you’ll find

.IsaacLab\source\standalone\workflows\rsl_rl\logs\rsl_rl\franka_reach

image
image1315×427 13.4 KB

This has the trained model_50.pt file or however many iterations you did

I had to play about with it to get the correct path but use this style of format for the play.py I had to add the load run with my folder in this case 2024-12-09_07-18-37 then the check point argument then the model.

if you have completed your full run you might not need the --checkpoint argument

python play.py --task Isaac-Reach-Franka-v0 --load_run 2024-12-09_07-18-37 --checkpoint model_50.pt

I got it to run on myside so hopefully this works for you. Let me know!