arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2605.20150cs.CVcs.PF
收藏

TideGS: 通过外存优化训练超过十亿个3D高斯溅射基元

TideGS: Scalable Training of Over One Billion 3D Gaussian Splatting Primitives via Out-of-Core Optimization

Hong Kong University of Science and Technology(香港科技大学);Tsinghua University(清华大学);Beijing Academy of Artificial Intelligence(北京人工智能研究院)

Chonghao Zhong, Linfeng Shi, Hua Chen, Tiecheng Sun, Hao Zhao, Binhang Yuan, Chaojian Li

AI总结:

针对大规模3D高斯溅射训练的内存瓶颈,提出TideGS外存训练框架,通过SSD-CPU-GPU层次化管理和三种协同技术,在单GPU上实现超过十亿高斯基元的训练并达到最优重建质量。

摘要:

训练十亿基元规模的3D高斯溅射(3DGS)本质上是内存受限的:每个高斯基元携带一个大的属性向量,总参数表迅速超出GPU容量,限制了先前系统在商用单GPU硬件上只能处理数千万高斯基元。我们观察到3DGS训练本质上是稀疏且轨迹条件的:每次迭代仅激活当前相机批次可见的高斯基元,因此GPU内存可以作为工作集缓存而非持久参数存储。基于这一洞察,我们引入了TideGS,一个外存训练框架,通过三种协同技术管理SSD-CPU-GPU层次结构中的参数:用于SSD对齐空间局部性的块虚拟化几何、用于将I/O与计算重叠的分层异步流水线,以及轨迹自适应差分流,该流在迭代之间仅传输增量工作集变化。实验表明,TideGS能够在单个24 GB GPU上训练超过十亿个高斯基元,同时在大规模场景中实现评估的单GPU基线中最佳的重建质量,超越了先前的外存基线(例如约1亿高斯基元)和标准内存训练(例如约1100万高斯基元)。

英文摘要:

Training 3D Gaussian Splatting (3DGS) at billion-primitive scale is fundamentally memory-bound: each Gaussian primitive carries a large attribute vector, and the aggregate parameter table quickly exceeds GPU capacity, limiting prior systems to tens of millions of Gaussians on commodity single-GPU hardware. We observe that 3DGS training is inherently sparse and trajectory-conditioned: each iteration activates only the Gaussians visible from the current camera batch, so GPU memory can serve as a working-set cache rather than a persistent parameter store. Building on this insight, we introduce TideGS, an out-of-core training framework that manages parameters across an SSD-CPU-GPU hierarchy via three synergistic techniques: block-virtualized geometry for SSD-aligned spatial locality, a hierarchical asynchronous pipeline to overlap I/O with computation, and trajectory-adaptive differential streaming that transfers only incremental working-set deltas between iterations. Experiments show that TideGS enables training with over one billion Gaussians on a single 24 GB GPU while achieving the best reconstruction quality among evaluated single-GPU baselines on large-scale scenes, scaling beyond prior out-of-core baselines (e.g., approximately 100M Gaussians) and standard in-memory training (e.g., approximately 11M Gaussians).

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑