How to visualize it in mujoco?

[JUNLIN1111]

How to visualize it in mujoco?

[nzolman]

Hi @JUNLIN1111—it depends on what you're looking to do here. In terms of visualizing a Mujoco environment from dm_control, I have primarily been using RLlib's dm_control wrapper (using ray 2.6.3) [1] and requesting visual observations from the environment through the render function [2]. At a low level, this call's dm_control's physics.render function [3].

from ray.rllib.env.wrappers.dm_control_wrapper import DMCEnv

# create env and request 480 x 480 resolution gif.
env = DMCEnv('cartpole', task_name='swingup',height=480, width=480)
env.reset()

# take the null action
obs = env.step(0.0 * env.action_space.sample())

# camera ID can either be 0 or 1, depending on the view you want.
cam_id = 0
pixels = env.render(camera_id=cam_id))

Does that meet your needs?

References
[1] https://github.com/ray-project/ray/blob/ray-2.6.3/rllib/env/wrappers/dm_control_wrapper.py#L77
[2] https://github.com/ray-project/ray/blob/ray-2.6.3/rllib/env/wrappers/dm_control_wrapper.py#L215
[3] https://github.com/google-deepmind/dm_control/blob/8358f6d605ce239e889851079862accc286c37e1/dm_control/mujoco/engine.py#L178

[Vcbby]

Hi there, I am trying to write a script to visualise the cart pole swing task under cart sparse policy, but I am a little bit confused and error occurs when I am coding. Can you show me a simple script to visualise the task pls?

[JUNLIN1111]

ber At 2025-01-06 02:18:13, "Vcbby" ***@***.***> wrote: Hi there, I am trying to write a script to visualise the cart pole swing task under cart sparse policy, but I am a little bit confused and error occurs when I am coding. Can you show me a simple script to visualise the task pls? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>