1
VOLD: Reasoning Transfer from LLMs to Vision-Language Models via On-Policy Distillation
利用同策略蒸馏将大语言模型的推理能力高效迁移至视觉语言模型，显著提升图文理解与推理表现。
arXiv:2510.23497v3 Announce Type: replace Abstract: Training vision-language models (VLMs) for complex reasoning remains a challenging task, i.a. due …