Robosuite can run with GLEW, but Robomimic can not.

[XizoB]

Sorry to raise some of my confusion with you. I was able to install robomimic with robosuite before and it was working, but recently I got an error reinstalling robomimic. The strange thing is that I can call GLEW with robosuite but not with robomimic !

Robosuite-v.1.2.0 Robomimic-v0.2.0

============= Initialized Observation Utils with Obs Spec =============

using obs modality: low_dim with keys: ['robot0_eef_pos', 'object', 'robot0_gripper_qpos', 'robot0_eef_quat'] using obs modality: rgb with keys: [] using obs modality: depth with keys: [] using obs modality: scan with keys: []

============= Loaded Environment Metadata ============= obs key object with shape (10,) obs key robot0_eef_pos with shape (3,) obs key robot0_eef_quat with shape (4,) obs key robot0_gripper_qpos with shape (2,) ERROR: GLEW initalization error: Missing GL version

[XizoB]

What's strange is that after I successfully ran the first robomimc, running it in a docker container didn't work, and then going back to the local side and running it again gave me an error (ERROR: GLEW initalization error: Missing GL version), which ended up causing both the local and docker containers to have errors

[amandlek]

This might be a mujoco-py issue - I believe robosuite v1.2 was still using mujoco-py instead of the DeepMind mujoco pip package.