JoVox: Joint Hierarchical Flow Matching for Lightweight Voxel Generation
Paper · https://arxiv.org/abs/2611.10690
Generating high-fidelity voxels while keeping their lightweight nature remains an open problem in computer vision. Current generators are either high-fidelity but computationally heavy, or lightweight but of lower fidelity. Hierarchical methods keep the lightweight nature of voxels. However, their coarse levels, which decide where all finer cells can appear, are generated with the least information about the shape and are never revisited, so they are often plausible but wrong. Our key insight is to let finer cells take part in deciding coarse cells. To this end, we propose JoVox, a discrete flow matching framework that learns the joint distribution of all levels of the voxel hierarchy, instead of a separate distribution for each level, so that coarse cells can be predicted with finer cells already generated nearby. Within this framework, we further introduce a two-level prediction, which outputs the cells two levels finer at once instead of one level finer, halving the number of levels that the network processes and yielding voxels of higher fidelity. We also introduce an unmasking strategy that revisits early decisions once finer cells are generated and corrects inconsistencies in masked diffusion models. Our framework outperforms the baselines on both unconditional and image-conditioned voxel generation. Compared with the current state-of-the-art TRELLIS.2, we achieve 8x higher training throughput, 1/15 of the FLOPs in a forward pass, train on 4 vs. 32 GPUs, and obtain higher quality with an 11% lower Chamfer distance.
First seen by Oriel on .
AI summary
JoVox is a discrete flow matching framework for voxel generation that jointly models hierarchical levels to improve fidelity while retaining lightweight representations.
- Jointly modeling the voxel hierarchy lets finer generated cells inform predictions of coarse cells, which conventional hierarchical methods do not revisit.
- Two-level prediction generates cells two levels finer at once, halving the number of hierarchy levels processed by the network.
- An unmasking strategy revisits early decisions after finer cells are generated to correct inconsistencies in masked diffusion models.
- Against TRELLIS.2, JoVox reports 8x higher training throughput, 1/15 the forward-pass FLOPs, training on 4 instead of 32 GPUs, and an 11% lower Chamfer distance.
- The framework outperforms baselines on both unconditional and image-conditioned voxel generation.
#voxel generation #flow matching #computer vision
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Hugging Face Papers: 4, latest observation ; +1 over observed period since
Comments
Comment count reported by the source.
- Hugging Face Papers: 2, latest observation ; +1 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hugging Face Papers | 2611.10690 | https://arxiv.org/abs/2611.10690 Read discussion |