arXiv · Machine Learning· Qize Yu, Lianrui Fan, Bowen Ping, Xini Ding, Zetian Song, Junbo Niu, Kaixuan Wang, Tianxing Chen, Yue Chen, Minghua He, Yuran Wang, Jie Huang, Haojun Zhang, Min Chen, Hao Li, Wenxuan Song, Ruihai Wu, Xianming Liu, Shilong Liu, Shuchang Zhou, Ping Luo, Shiyu Huang·· 17 小时前AI 评分57
GroundAnything：以并行解码实现快速且精准的视觉定位
GroundAnything: Reconciling Parallel Decoding with Precise Visual Grounding at Flash Speed
AI 导读
GroundAnything 是一个 4B 参数的视觉定位基础模型，通过 blockwise denoising 将空间预测并行化，并结合 AR 转扩散训练、监督微调和 GRPO 强化后训练。
来源：arXiv · Machine Learning · arxiv.org