Take Mugs Off Of Shelf¶
Agent runs
Each mode of this task runs once per run. Est. cost is tokens at list price (data/prices.yml), never a bill; see the RoboLab runs for every task of a run.
Codex + GPT-6 Luna, reasoning xhigh (ChatGPT login)default run
| Trial | Started | Agent time | Model requests | Tokens in / out | Est. cost | Grade |
|---|---|---|---|---|---|---|
| U✓ 1h 11m | 09-26 14:02 | 1h 11m | 201 | 25.3M / 101k | $0.334 | deterministic 1, grader_error 0, instance_exact 1, instance_ok 1, n_actions 321, video_rendered 1 |
| L✗ 1h 30m | 09-29 06:46 | 1h 30m | 189 | 21.5M / 196k | $0.377 | deterministic 1, grader_error 0, live_success 0, n_actions 1726, replay_success 0, video_rendered 1 |
Task instruction (upstream)
Take the mugs off the shelf
No oracle video published upstream — this is the scene it starts from.
What RoboLab states about this task
| Success predicate | object_outside_of_and_on_surface |
| Predicate arguments | object: ['ceramic_mug', 'mug'] · container: rack_l04 · surface: table · logical: all · require_gripper_detached: True |
| Scored subtasks | 1 |
| Subtask predicates | pick_and_place_on_surface |
| Objects | ceramic_mug mug rack_l04 serving_bowl utilityjug_a01 |
| Upstream attributes | semantics affordance |
| Upstream difficulty | simple |
| Episode budget | 180 s |
| Other wordings | vague — Clear the shelf specific — Remove each mug from the second level of the shelf and place them on the table below |
| Environment class | TakeMugsOffOfShelfTask |
| Upstream name | TakeMugsOffOfShelfTask |
| Defined in | robolab/tasks/benchmark/take_mugs_off_of_shelf.py |
From https://github.com/NVlabs/RoboLab @ v0.3.1.
Tags¶
Task DomainManipulation
Skill primitives in the demo: affordancesemantics
Why this task is interesting¶
Not yet written.
Capability notes¶
Not yet written.
Oracle demo review¶
Not yet reviewed.