The authors work with VecSet-based single-stage 3D generative models, which represent shapes as unordered sets of latent tokens.
Position-based self-conditioning improves single-stage 3D generation without extra stages
The method recovers token positions from latent representations and provides coarse-to-fine spatial guidance during denoising.
Independent
Ziheng Ouyang · Zeqiang Lai · Jiarui Chen · Jiangshan Wang · Yuhao Wan · Jingbo Gong · +4 more
Research Digest··2 min read
The authors propose Position Forcing, a self-conditioning framework that recovers token positions from predicted clean latents, quantizes them at progressively finer resolutions according to the noise level, and feeds the positional encodings back into a diffusion Transformer.
Why this paper
Independent
In one line
Position Forcing improves single-stage 3D generation by using recovered token positions as progressive spatial guidance during denoising.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§