TensorX
返回文献探索

Paper · arXiv 2509.15212

RynnVLA-001: Using Human Demonstrations to Improve Robot Manipulation

Yuming Jiang, Siteng Huang, Shengke Xue, Yaxi Zhao, Jun Cen, Sicong Leng, Kehan Li, Jiayan Guo, Kexiang Wang, Mingxiu Chen, Fan Wang, Deli Zhao, Xin Li

22 upvotesSeptember 18, 2025arXiv 预印本
AI 摘要

RynnVLA-001, a vision-language-action model, uses a two-stage pretraining approach and ActionVAE to achieve superior performance on robotics tasks.

Ego-Centric Video Generative PretrainingHuman-Centric Trajectory-Aware ModelingImage-to-Video modelkeypoint trajectoriesActionVAEvariational autoencoderlatent embeddingsVLA model

Abstract

This paper presents RynnVLA-001, a vision-language-action(VLA) model built upon large-scale video generative pretraining from human demonstrations. We propose a novel two-stage pretraining methodology. The first stage, Ego-Centric Video Generative Pretraining, trains an Image-to-Video model on 12M ego-centric manipulation videos to predict future frames conditioned on an initial frame and a language instruction. The second stage, Human-Centric Trajectory-Aware Modeling, extends this by jointly predicting future keypoint trajectories, thereby effectively bridging visual frame prediction with action prediction. Furthermore, to enhance action representation, we propose ActionVAE, a variational autoencoder that compresses sequences of actions into compact latent embeddings, reducing the complexity of the VLA output space. When finetuned on the same downstream robotics datasets, RynnVLA-001 achieves superior performance over state-of-the-art baselines, demonstrating that the proposed pretraining strategy provides a more effective initialization for VLA models.

北京市昌平区探索星信息技术及软件开发工作室

京ICP备2026059466号