Finetuning VLAs for Dexterous Manipulation: The Right Checkpoint Is All You Need
Vision-Language-Action models have demonstrated strong manipulation capabilities. A limitation of VLA models is that most manipulation datasets focus on parallel-jaw gripper systems, making transfer to multi-finger hands challenging, particularly as teleoperation datasets are harder to collect with multi-fingered robots. In this work, we study how to best adapt VLAs to dexterous hands by pairing supervised fine-tuning (SFT) with diffusion-steering reinforcement learning (DSRL). Using $π_{0.5}$ and a LEAP hand, we evaluate five fine-tuning recipes across two simulation tasks using only a handful of demonstrations. Our empirical analysis yields three key insights: (1) SFT success is poorly correlated with post-DSRL success, meaning the best SFT checkpoint is rarely the most effective to steer; (2) freezing the vision-language model (VLM) and tuning only the action expert produces the most steerable policy for both training and out-of-distribution placements; and (3) low-rank adaptation (LoRA) is the least steerable approach. Consequently, we recommend an adaptation recipe that freezes the VLM, fully trains the action expert, and eschews SFT rollout success as a checkpoint selection metric.
Authors
- Kirsty Ellis
- Glen Berseth (ORCID: https://orcid.org/0000-0001-7351-8028)
- Michael Przystupa
- Charlotte Cloutier (ORCID: https://orcid.org/0000-0001-5633-5668)
- Artur Kuramshin
- Mohammad S. Akhter
Publication Details
- Journal
- arXiv (Cornell University)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.48550/arxiv.2611.23676
- Primary Topic
- Robot Manipulation and Learning
- Type
- preprint