Kelly Guo

Hi there, this seems to be an error related to the visual materials in the asset and it shouldn't influence the final results of training. You can also verify by...

Hi there, unfortunately this feature is not currently available in OmniIsaacGymEnvs. We will look into adding this in the future.

Hi @AlanSunHR , Your setup looks correct, you can try verifying whether the parameters are being randomized by calling `knee_view._physics_view.get_material_properties()`. Values in the UI will not be modified when using...

Hi there, please make sure you are running the `pip install -e .` command from within the OmniIsaacGymEnvs folder.

Hi there, it should be possible to use WebRTC and WebSocket Browser Clients by adding the following dependencies to the `omni.isaac.sim.python.gym.headless.render.kit` app file: ``` "omni.services.streamclient.websocket" = {} "omni.services.streamclient.webrtc" = {}...

Hi there, thanks for the report. We will look into this.

Hi there, there are a few parameters that you can try tuning to achieve more accurate velocities. Decreasing the simulation dt can help with numerical issues when the velocities converge,...

You can try setting velocity iteration to 0, since the velocity iterations will tamper with the velocities, but those updates won't be reflected in the positions. Our physics team is...

Hi there, can you try lowering the number of environments with the `num_envs` flag? It's possible the machine ran out of memory during the deformable simulation.

Hi there, we do not currently provide APIs for getting accelerations, but they can be approximated using the velocities.