loading page

Hash3D: Training-free Acceleration for 3D Generation
  • Xingyi Yang,
  • Xinchao Wang
Xingyi Yang
National University of Singapore

Corresponding Author:[email protected]

Author Profile
Xinchao Wang
National University of Singapore


The evolution of 3D generative modeling has been notably propelled by the adoption of 2D diffusion models. Despite this progress, the cumbersome optimization process per se presents a critical hurdle to efficiency. In this paper, we introduce Hash3D, a universal acceleration for 3D generation without model training. Central to Hash3D is the insight that feature-map redundancy is prevalent in images rendered from camera positions and diffusion time-steps in close proximity. By effectively hashing and reusing these feature maps across neighboring timesteps and camera angles, Hash3D substantially prevents redundant calculations, thus accelerating the diffusion model's inference in 3D generation tasks. We achieve this through an adaptive grid-based hashing. Surprisingly, this feature-sharing mechanism not only speed up the generation but also enhances the smoothness and view consistency of the synthesized 3D objects. Our experiments covering 5 textto-3D and 3 image-to-3D models, demonstrate Hash3D's versatility to speed up optimization, enhancing efficiency by 1.3 ∼ 4×. Additionally, Hash3D's integration with 3D Gaussian splatting largely speeds up 3D model creation, reducing text-to-3D processing to about 10 minutes and image-to-3D conversion to roughly 30 seconds. The code is provided in https://github.com/Adamdad/hash3D.
02 Apr 2024Submitted to TechRxiv
02 Apr 2024Published in TechRxiv