I did not find the fetch on the github web

[beforeface]

How could I replace these files ? I did not find these files

[hyy02]

Sorry for that, the simulation is redeveloping, the documentation doesn't keep consistent with the repo update right now, After the simulation benchmark developing finished, the documentaion will update, for now You can change the file like this:

XLerobot/simulation/Maniskill/agents/xlerobot -> ManiSkill/mani_skill/agents/robots/xlerobot
XLerobot/simulation/Maniskill/assets/xlerobot-> ManiSkill/mani_skill/assets/robots/xlerobot

and see the README in simulation part.

[beforeface]

Yes, thank you. That did actually solved my problem. But I still has some questions.

1. from the README, I know you update the command running the example. But I tried to run a double-arm version of XLeRobot, I got a message " python -m mani_skill.examples.demo_ctrl_action_ee_keyboard -e "ReplicaCAD_SceneManipulation-v1" -r "xlerobot" --render-mode="human" --shader="rt-fast" -c "pd_joint_delta_pos" This is my command "python -m mani_skill.examples.demo_ctrl_action_ee_keyboard -e "ReplicaCAD_SceneManipulation-v1" -r "xlerobot" --render-mode="human" --shader="rt-fast" -c "pd_joint_delta_pos". This is what I do. XLerobot/simulation/Maniskill/agents/xlerobot -> ManiSkill/mani_skill/agents/robots/xlerobot XLerobot/simulation/Maniskill/assets/xlerobot-> ManiSkill/mani_skill/assets/robots/xlerobot XLerobot/simulation/Maniskill/examples -> ManiSkill/mani_skill/exaples XLerobot/simulation/Maniskill/envs/Scenes-> ManiSkill/mani_skill/envs

if self.env.robot_uids == "fetch" or self.env.robot_uids == "xlerobot_single" or self.env.robot_uids == "xlerobot": self.env.agent.reset(self.env.agent.keyframes["rest"].qpos) self.env.agent.robot.set_pose(sapien.Pose([-1, 0, 0.02])) else: raise NotImplementedError(self.env.robot_uids)

2. From XLeRobot the page, I see lots of new scripts, how could I do with these?

3. From XLeRobot this page, I need to copy all the codes of XLeRobot config files to the corresponding place of lerobot. I'm a little confused about "the corresponding place of lerobot". What's the "corresponding place of lerobot"?

[hyy02]

1. For the question 1, I try the author's code, it's just a warning, you can change the code here to SUPPORTED_ROBOTS = ["panda", "fetch", "xlerobot"] , and make sure you change the scene_builder and init described in the author's README.
2. For the question 2, these are the scripts for real world lerobot control, it's not for the maniskill simulation. If you want to run simulation for it please see the simulation part.
3. For the question 3

model -> leorbot/src/lerobot/model
robots/xlerobot ->lerobot/src/lerobot/robots/xlerobot

@beforeface

[Vector-Wangel]

Thanks for the questions and answering, I will update the documentation today to address these concerns.

[beforeface]

Okay ,thank you.

1. Have you successfully run the double-arm version? I tried to add the "xlerobot" uid, still got a problem
2. I bought some materials from web, meanwhile I tried to understand the tutorial and found the issue.
3. Maybe the files for the real world XLerobot? I need to mkdir the src files?

[hyy02]

@beforeface

1. Can you show your terminal output, so I can know where the bug is?
2. Good job, welcome to join the XLerobot family
3. You needn't, you just need to clone the lerobot repo

[beforeface]

1. This is what I did in the code

This is the command and output. It looks like that tried to access an array but out of bound.

2. I'm really happy. This is my first real robot project.

3. Okay, thank your reply. I hope I can run it in the real world.

[hyy02]

@beforeface It seems that you choose the wrong controller, end effector control need the pd_joint_delta_pos_dual_arm, please use the command in the documentation. python -m mani_skill.examples.demo_ctrl_action -e "ReplicaCAD_SceneManipulation-v1" --render-mode="human" --shader="rt-fast" -c "pd_joint_delta_pos_dual_arm"

[beforeface]

@hyy02 No, it still not work for it. the message is "pd_joint_delta_pos_dual_arm not in supported modes". Actually I'm a new guy for the mani-skill, it looks like I need to edit the mani-skill's config?

[hyy02]

@beforeface Sorry :joy: , I forgot to add -r "xlerobot", the default agent setting is 'fetch', I test this: python -m mani_skill.examples.demo_ctrl_action -e "ReplicaCAD_SceneManipulation-v1" --render-mode="human" --shader="rt-fast" -c "pd_joint_delta_pos_dual_arm" -r "xlerobot"

[beforeface]

@hyy02 Yeah, thank you. The double-arm version successfully run. I have a question: how could you know the -c parameter is wrong?

[hyy02]

@beforeface Actually, I don't really understand your question 😂 , Just follow the author's documentation, use the controller config he specified. the single-arm and dual-arm use different controller config, checkout the terminal output with the author's documentation

[Vector-Wangel]

@hyy02 Yeah, thank you. The double-arm version successfully run. I have a question: how could you know the -c parameter is wrong?

If you mean how to debug errors like "pd_joint_delta_pos_dual_arm not in supported modes", the control config is usually defined in the robot config file, in this case "/agents/robots/xlerobot/xlerobot.py". You were previously having this error because xlerobot robot file wasn't loaded correctly. Later if you want to add or modify controller configs you can just do it in the robot config file. You can check Maniskill Docs and tutorials for more details.

[beforeface]

@Vector-Wangel Thank your replay. I think I need take time to read Maniskill Docs.