VLA-OPD: Bridging Offline SFT and Online RL for Vision-Language-Action Models via On-Policy Distillation

@misc{zhong2026vlaopd,
  title = {VLA-OPD: Bridging Offline SFT and Online RL for Vision-Language-Action Models via On-Policy Distillation},
  author = {Zhide Zhong and Haodong Yan and Junfeng Li and Junjie He and Tianran Zhang and Haoang Li},
  year = {2026},
  eprint = {2603.26666},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2603.26666},
}