FastOPD: On-Policy Distillation for Lightweight VLA Deployment

Hugging Face Daily Papers Papers
vla on-policy-distillation robotics vision-language-action efficiency model-compression flow-based-policy

Summary

FastOPD proposes an on-policy distillation framework that compresses large Vision-Language-Action (VLA) foundation models into lightweight students with few inference steps, retaining 84% of π₀.₅'s LIBERO performance at a 78.1% latency reduction and deploying a compact student distilled from MolmoAct2 on a real robot.

Vision-Language-Action (VLA) foundation models have scaled rapidly to enhance manipulation performance and generalizability, but this scaling incurs high computational costs that render real-world deployment increasingly challenging. Existing approaches typically mitigate this issue by designing smaller architectures or reducing the iterative denoising steps in flow-based policies. In this work, we propose FastOPD, a foundation-to-lightweight VLA framework that enables the practical deployment of large-scale VLAs through efficient on-policy distillation. Specifically, FastOPD adapts a flow map for single-state teacher supervision and combines it with a self-consistency objective to construct a compact student that learns the teacher dynamics. Furthermore, we theoretically demonstrate that minimizing this objective allows the distilled student to recover a distribution on par with that induced by an ideal few-step teacher model. We evaluate FastOPD across diverse foundation policies in simulation and real-world experiments. On LIBERO, FastOPD retains 84% of the performance of π_{0.5} with only two inference steps, reducing inference latency by 78.1% while outperforming existing few-step distillation baselines in average success rate. With LingBot-VLA as the teacher, FastOPD improves the single-step success rate over the base student by 15.9 percentage points on RoboTwin 2.0. We further demonstrate its applicability to a World Action Model (WAM) and deploy a compact student distilled from MolmoAct2 on a real robot.
Original Article
View Cached Full Text

Cached at: 10/08/26, 12:11 PM

Paper page - FastOPD: On-Policy Distillation for Lightweight VLA Deployment

Source: https://huggingface.co/papers/2610.02832

Abstract

Vision-Language-Action(VLA)foundationmodelshavescaledrapidlytoenhancemanipulationperformanceandgeneralizability,butthisscalingincurshighcomputationalcoststhatrenderreal-worlddeploymentincreasinglychallenging.Existingapproachestypicallymitigatethisissuebydesigningsmallerarchitecturesorreducingtheiterativedenoisingstepsinflow-basedpolicies.Inthiswork,weproposeFastOPD,afoundation-to-lightweightVLAframeworkthatenablesthepracticaldeploymentoflarge-scaleVLAsthroughefficienton-policydistillation.Specifically,FastOPDadaptsaflowmapforsingle-stateteachersupervisionandcombinesitwithaself-consistencyobjectivetoconstructacompactstudentthatlearnstheteacherdynamics.Furthermore,wetheoreticallydemonstratethatminimizingthisobjectiveallowsthedistilledstudenttorecoveradistributiononparwiththatinducedbyanidealfew-stepteachermodel.WeevaluateFastOPDacrossdiversefoundationpoliciesinsimulationandreal-worldexperiments.OnLIBERO,FastOPDretains84%oftheperformanceofπ_{0.5}withonlytwoinferencesteps,reducinginferencelatencyby78.1%whileoutperformingexistingfew-stepdistillationbaselinesinaveragesuccessrate.WithLingBot-VLAastheteacher,FastOPDimprovesthesingle-stepsuccessrateoverthebasestudentby15.9percentagepointsonRoboTwin2.0.WefurtherdemonstrateitsapplicabilitytoaWorldActionModel(WAM)anddeployacompactstudentdistilledfromMolmoAct2onarealrobot.

View arXiv pageView PDFProject pageAdd to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.02832 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.02832 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.02832 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

ShortOPD: Recovering Pruned LLMs with Short-to-Long On-Policy Distillation

arXiv cs.LG

ShortOPD proposes a short-to-long on-policy distillation schedule that recovers pruned LLMs for free-form generation by focusing training on effective prefixes, achieving up to 9x improvement over unrecovered models and matching long-horizon distillation with a quarter of the training time.

KL for a KL: On-Policy Distillation with Control Variate Baseline

Hugging Face Daily Papers

Proposes vOPD, which stabilizes on-policy distillation for LLMs by introducing a control variate baseline from reinforcement learning, achieving performance comparable to expensive full-vocabulary methods at lower computational cost.

OPD-V: Visual On-Policy Self-Distillation with Modality Balance

Hugging Face Daily Papers

Introduces OPD-V, a visual on-policy self-distillation paradigm for multimodal large language models that leverages positive and negative teachers to exploit modality balance as privileged information, improving reasoning performance across benchmarks while reducing training cost.

LastOPD: Taming Collapse in Latent On-Policy Distillation

arXiv cs.LG

The paper proposes LastOPD, a method to prevent collapse in latent on-policy distillation by applying latent signals only at the last layer during a short crossfade period, leading to improved performance on benchmarks like MATH-500.