Conversation
|
we do need the same change for |
good point, ill update mink as well |
| all_robot_action_dicts.append(action_dict) | ||
| continue | ||
|
|
||
| inactive_robot_ac_dict = copy.deepcopy(all_prev_gripper_actions[i]) |
There was a problem hiding this comment.
Setting inactive robot action as prev gripper action is confusing to me, as in active robot action seems more than just gripper. Can we rename things to be more intuitive?
There was a problem hiding this comment.
Yeah good points my bad, Ill refactor accordingly
| raise NotImplementedError | ||
|
|
||
|
|
||
| def get_abs_arm_static_target(robot, arm, env): |
There was a problem hiding this comment.
why's the fn called get_abs_arm_static_target? what's an absolute arm? does it mean arm's controller is absolute and doesn't work for relative action space? Can the function just be to get_arm_target() or get_arm_current_target?
|
|
||
| def get_arm_ref(env, robot, arm): | ||
| """ | ||
| Extracts the reference pose of the specified arm of the robot. |
There was a problem hiding this comment.
what's reference pose? the world pose target? docs might be confusing to users
2 participants
What this does
Fixes #737
It specifically fixes two issues when initializing two robots.
How it was tested
Run any environment with two robots with absolute action control before and after these changes were made for example
python robosuite/scripts/collect_human_demonstrations.py --environment TwoArmLift --device spacemouse --robots Panda Panda --config parallel --controller WHOLE_BODY_IKBefore the changes the second robot will never stay in the same pose when not being controller After these changes they do