chyun/thinkworld_final_oven_robot_human_composition_v1
ThinkWorld oven composition prompt v1 This is an additive metadata derivation of chyun/thinkworld_final_oven_robot_human@a0c9dcca7a9e2ba394c306bcdf6e7161b712dbc0. Video bytes, timing, robot state/action labels, supervision masks, camera layout, and source provenance are unchanged. Human episodes retain their original global composite prompt exactly. The ordinary LeRobot task resolves from task_index to the episode-global instruction. atomic_task_index retains the current skill… See the full description on the dataset page: https://huggingface.co/datasets/chyun/thinkworld_final_oven_robot_human_composition_v1.
ThinkWorld oven composition prompt v1
This is an additive metadata derivation of chyun/thinkworld_final_oven_robot_human@a0c9dcca7a9e2ba394c306bcdf6e7161b712dbc0. Video bytes, timing, robot state/action labels, supervision masks, camera layout, and source provenance are unchanged. Human episodes retain their original global composite prompt exactly.
The ordinary LeRobot task resolves from task_index to the episode-global instruction. atomic_task_index retains the current skill, and allow_skill_dropout gates replacement of the entire current-skill clause by the fixed inference clause. The parent dataset is not modified.
Evaluation episodes: [1, 3, 12, 14, 21, 31, 32, 36, 40, 46, 50, 53, 63, 66, 73, 78, 79, 91, 93, 98, 124, 125, 126, 127, 128, 129, 146, 147, 148, 149, 166, 167, 168, 169, 186, 187, 188, 189]. The authoritative split and full derivation contract are in meta/splits.json and meta/manifest.json.
