Feedback from humans

[doxav]

According to your paper, you used human feedback in Voyager to achieve new tasks like complex 3D structures.

How to reproduce the experiment ? Test both human as a critic and human as a curriculum ?

[doxav]

@xieleo5 @theblazehen are the human feedback (mentioned in paper) in the github repo cause I could not find it ?

[xieleo5]

Hi, sorry for the late reply. I've been busy these days. For human in-loop experiments, you can follow this way to instantiate voyager:

voyager = Voyager(
    ...
    reset_placed_if_failed=True,
    action_agent_task_max_retries=100,
    action_agent_show_chat_log=False,
    curriculum_agent_mode="manual",
    critic_agent_mode="manual",
    ...
)

You will be asked to give the next task and judge whether the task is finished.

[github-actions]

This issue is stale because it has been open for 30 days with no activity.

[github-actions]

This issue was closed because it has been inactive for 14 days since being marked as stale.