openral/openral/omdet-turbo-indoor

omdet-turbo-indoor — rSkill discovery view

Generated view, not a hand-written skill. This SKILL.md is a discovery-only

mirror of `rskill.yaml`, produced by tools/generate_rskill_skillmd.py.

It lets tools that read the standard agent-skill format find and reason about this

OpenRAL rSkill. The rskill.yaml manifest is the single source of truth

(CLAUDE.md §1.3). Do not edit by hand — edit the manifest and regenerate.

What it is

An OpenRAL object detector (role: s1, kind: detector). OmDet-Turbo (Swin-tiny) real-time open-vocabulary detector, run in-process over a fixed curated indoor vocabulary (~230 household/kitchen/office/ manipulation classes). An unprompted background producer: publishes ObjectsMetadata to /openral/perception/objects every frame without reasoner prompting, giving the world model far more classes than the 80 COCO categories. Apache-2.0 weights from omlab/omdet-turbo-swin-tiny-hf. This is the zeroshot_hf detector engine.

Capabilities

• Verbs: detect
• Objects: kitchenware (cup, mug, bottle, bowl, plate, pot, pan, kettle, utensils) · food (fruit, vegetables, bread, packaged food) · appliances (fridge, microwave, oven, toaster, kettle, TV) · electronics (laptop, monitor, keyboard, phone, remote, charger, cables) · furniture (chair, sofa, table, desk, bed, shelf, cabinet, drawer) · bathroom items (toilet, sink, towel, soap, toothbrush, mirror) · office and stationery (book, pen, scissors, stapler, folder, notebook) · cleaning and laundry (broom, mop, bucket, sponge, spray bottle, hamper) · containers and personal items (box, basket, bag, backpack, keys, glasses) · tools and hardware (hammer, screwdriver, drill, switch, outlet, door handle) · clothing and wearables (shoe, hat, glove, jacket) · toys (teddy bear, doll, ball, blocks)
• Scenes: tabletop · kitchen · indoor · household · office · bathroom
• Embodiments: any

Why this is discovery-only

An agent skill is natural-language instructions loaded into an LLM's context. An rSkill is an executable artifact: it carries a typed capability/embodiment contract, model weights, a runtime, and a license/provenance gate — none of which fit in freeform markdown. So an agent can use this view to select the right skill, but cannot execute it by loading this file. Execution always goes through the OpenRAL loader and the robot HAL.

License

• Code: Apache-2.0.
• Weights: apache-2.0 — permissive / commercial-use OK

How to actually run it (not via an agent harness)

from openral_rskill import rSkill
skill = rSkill.from_pretrained("OpenRAL/rskill-omdet_turbo-any-indoor-fp16")
# the loader validates embodiment / sensors / runtime / quantization against the target
# RobotDescription and enforces the weight-license gate before any weights load.

See `rskill.yaml` for the authoritative, validated manifest.