Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

VirtualHome

Benchmarks

Task NameDataset NameSOTA ResultTrend
Human-robot teaming for household object rearrangementVirtualHome 2.0 (test)
Success Rate80
19
Robotic Task Planning in Dynamic EnvironmentsVirtualHome
Success Rate92
16
Interactive decision makingVirtualhome
Success Rate59.1
15
Instruction ExecutionVirtualHome Unseen domains
Success Rate83.61
15
Continual Instruction FollowingVirtualHome
SR61.12
15
Embodied Task PlanningVirtualHome (Seen)
Simple Success9,140
10
Multi-agent coordinationVirtualHome-Social Addition 4→5
Average Steps48.213
6
Embodied Task PlanningVirtualHome (unseen domains)
Success Rate80.16
6
Few-shot task expansionVirtualHome average performance (unseen domains)
SR82.59
5
Few-shot task expansionVirtualHome unseen domains 5-shot
Success Rate83.61
5
Few-shot task expansionVirtualHome unseen domains 1-shot
SR81.56
5
Embodied Task PlanningVirtualHome Novel Apartment (Unseen)
Simple Success Rate82.9
4
SetuptableVirtualHome livingroom_and_bedroom
Task Success Rate (TSR)82.7
3
SetuptableVirtualHome kitchen_and_bedroom
TSR81.9
3
PutfridgeVirtualHome kitchen and bathroom
TSR80.6
3
PutfridgeVirtualHome bathroom_and_livingroom
TSR81.9
3
PreparefoodVirtualHome bedroom_and_bathroom
TSR80.8
3
PreparefoodVirtualHome kitchen and livingroom
TSR80.8
3
PutdishwasherVirtualHome livingroom_and_bedroom
TSR82.8
3
ReadbookVirtualHome bedroom_and_bathroom
TSR83.1
3
ReadbookVirtualHome bedroom_and_kitchen
TSR85.4
3
Showing 21 of 21 rows