3D Gaussian Splatting has demonstrated notable success in large-scale scene reconstruction, but challenges persist due to high training memory consumption and storage overhead. Hybrid representations that integrate implicit and explicit features offer a way to mitigate these limitations. However, when applied in parallelized block-wise training, two critical issues arise since reconstruction accuracy deteriorates due to reduced data diversity when training each block independently, and parallel training restricts the number of divided blocks to the available number of GPUs. To address these issues, we propose Momentum-GS, a novel approach that leverages momentum-based self-distillation to promote consistency and accuracy across the blocks while decoupling the number of blocks from the physical GPU count. Our method maintains a teacher Gaussian decoder updated with momentum, ensuring a stable reference during training. This teacher provides each block with global guidance in a self-distillation manner, promoting spatial consistency in reconstruction. To further ensure consistency across the blocks, we incorporate block weighting, dynamically adjusting each block's weight according to its reconstruction accuracy. Extensive experiments on large-scale scenes show that our method consistently outperforms existing techniques, achieving a 12.8% improvement in LPIPS over CityGaussian with much fewer divided blocks and establishing a new state of the art. Project page: this https URL
3D高斯点云（3D Gaussian Splatting）在大规模场景重建中取得了显著成功，但由于训练过程中高内存消耗和存储开销，仍面临诸多挑战。结合隐式与显式特征的混合表示为缓解这些限制提供了可能。然而，在并行分块训练中应用时会出现两个关键问题：由于每个块独立训练导致数据多样性下降，重建精度随之恶化；并且并行训练限制了划分块的数量，受制于可用GPU的数量。 为了解决这些问题，我们提出了Momentum-GS，这是一种利用基于动量的自蒸馏方法的新颖框架，旨在提升各块之间的一致性和精度，同时将划分块的数量从物理GPU数量中解耦。我们的方法保持一个通过动量更新的教师高斯解码器，在训练过程中提供稳定的参考。该教师模型以自蒸馏的方式为每个块提供全局指导，促进空间一致性的重建。此外，为了进一步保证块间的一致性，我们引入了块权重机制，动态调整每个块的权重以匹配其重建精度。 在大规模场景上的广泛实验表明，我们的方法在保持较少划分块的同时，性能显著优于现有技术。在LPIPS指标上，相比CityGaussian提升了12.8%，并在领域内树立了新的技术标杆。