VLA-MOPD: On-Policy Distillation for Multi-Task VLA Policy Consolidation
Physical intelligence requires embodied agents to ground language in observations and execute actions while retaining useful skills. Vision-language-action (VLA) policies offer a shared interface, but task-specific post-training often produces separate specialists. Consolidation is difficult: mixed-task reinforcement learning provides sparse trajectory-level credit, while offline distillation may miss student-visited observations and endpoint distillation does not supervise intermediate action-generation states. We introduce VLA Multi-teacher On-Policy Distillation (VLA-MOPD), a protocol for consolidating flow-matching VLA specialists. After acquiring task or suite experts, a shared student collects robot rollouts and queries the instruction-routed frozen teacher at each visited observation and intermediate ordinary differential equation (ODE) state. Detached per-flow-step transition targets supervise valid action chunks through a masked local matching loss; an optional reference-policy term supports retention on other tasks. Starting from one trajectory per LIBERO task, VLA-MOPD raises mean success-once from 75.4% to 98.9%, compared with 98.15% for endpoint action distillation.…
Multi-task VLA policy consolidation via on-policy distillation
- arXiv
- 2611.12286
- Authors
- Yongxuan Lv, Zhen Fang, Lichen Man, Hao Li, Pu Li and 8 more
- Categories
- cs.RO · cs.CV