We present V-JEPA 2.1, a family of self-supervised models that learn dense, high-quality visual representations for both images and videos while retaining strong global scene understanding. The approach combines four key components. First, a dense predictive loss uses a masking-based objective in which both visible and masked tokens contribute to the training signal, encouraging explicit spatial and temporal grounding. Second, deep self-supervision applies the self-supervised objective hierarchically across multiple intermediate encoder layers to improve representation quality. Third, multi-modal tokenizers enable unified training across images and videos. Finally, the model benefits from effective scaling in both model capacity and training data. Together, these design choices produce representations that are spatially structured, semantically coherent, and temporally consistent. Empirically, V-JEPA 2.1 achieves state-of-the-art performance on several challenging benchmarks, including 7.71 mAP on Ego4D for short-term object-interaction anticipation and 40.8 Recall@5 on EPIC-KITCHENS for high-level action anticipation, as well as a 20-point improvement in real-robot grasping success rate over V-JEPA-2 AC. The model also demonstrates strong performance in robotic navigation (5.687 ATE on TartanDrive), depth estimation (0.307 RMSE on NYUv2 with a linear probe), and global recognition (77.7 on Something-Something-V2). These results show that V-JEPA 2.1 significantly advances the state of the art in dense visual understanding and world modeling.

翻译：我们提出V-JEPA 2.1，这是一系列自监督模型家族，能够在保持强大全局场景理解能力的同时，为图像和视频学习密集且高质量的视觉表征。该方法融合了四个关键组件。首先，密集预测损失采用基于掩码的目标函数，其中可见与掩码令牌均参与训练信号生成，从而促进显式的空间和时间锚定。其次，深度自监督机制在编码器多个中间层级上分层应用自监督目标，以提升表征质量。第三，多模态分词器支持图像与视频的统一训练。最后，模型得益于模型容量与训练数据的有效扩展。这些设计选择共同产生了具备空间结构化、语义一致性与时间连续性的表征。实验表明，V-JEPA 2.1在多项具有挑战性的基准测试中达到了最先进性能，包括Ego4D短程物体交互预测任务中7.71 mAP、EPIC-KITCHENS高层动作预测任务中40.8%的Recall@5，以及相比V-JEPA-2 AC在真实机器人抓取成功率上提升20个百分点。该模型在机器人导航（TartanDrive上5.687 ATE）、深度估计（NYUv2上线性探针评估0.307 RMSE）及全局识别（Something-Something-V2上77.7）方面同样展现出强劲性能。这些结果表明，V-JEPA 2.1显著推动了密集视觉理解与世界建模领域的技术发展。