AI 导读
我们发布 SimpleNav：一个面向 Navigation VLA 的统一研究框架。🤗 Navigation Vision-Language-Action（VLA）研究往往需要在不同数据集、模型和评估环境之间反复做工程适配。 为了让 Navigation VLA 开发更可复现、更高效，我们推出 SimpleNav，一个开源框架，统一了从数据处理、模型训练到闭环评估的完整流程。 👇
正文
We are releasing SimpleNav: a unified research framework for Navigation VLA. 🤗
Navigation Vision-Language-Action (VLA) research often requires repeated engineering efforts across different datasets, models, and evaluation environments.
To make Navigation VLA development more reproducible and efficient, we introduce SimpleNav, an open-source framework that unifies the full pipeline from data processing, model training to closed-loop evaluation.
👇
来源：@OpenBMB · x.com