Research map / 3D vision / 3D generation
Fast3Dcache: Training-free 3D Geometry Synthesis Acceleration
Mengyu Yang, Yanming Yang, Chenyi Xu, Chenxi Song, Yufan Zuo, Tong Zhao, Ruibo Li, Chi Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| apply 2D-style RAS caching to 3D | TRELLIS · Toys4K | F-Score: 54.8244 -> 40.2769 | Hurt |
| apply Fast3Dcache (tau=8) | TRELLIS · Toys4K | throughput (iter/s): 0.5055 -> 0.6426 | Improved |
| apply Fast3Dcache (tau=8) | TRELLIS · Toys4K | F-Score: 54.8244 -> 53.7528 | Hurt |
| replace fixed-ratio caching with PCSC scheduler (ablation) | Fast3Dcache · Toys4K | F-Score: 34.5122 -> 54.0900 | Improved |
| vary decay slope mu by an order of magnitude (ablation) | PCSC · Toys4K | F-Score: 53.4978 to 54.0900 | No clear effect |
| use velocity only for token selection (ablation) | SSC · Toys4K | F-Score: 53.9156 -> 44.9630 | Hurt |
| add acceleration to velocity criterion (joint) (ablation) | SSC · Toys4K | F-Score: 53.5394 -> 54.0900 | Improved |
| add periodic elimination step tau (ablation) | Fast3Dcache · Toys4K | F-Score: 51.8157 -> 53.7528 | Improved |
| combine Fast3Dcache with TeaCache | TeaCache on TRELLIS · Toys4K | throughput (iters/s): 1.45 -> 1.74 | Improved |
| combine Fast3Dcache with TeaCache | TeaCache on TRELLIS · Toys4K | F-Score: 53.5567 -> 53.9420 | Improved |
| increase tau from 3 to 8 (ablation) | Fast3Dcache · Toys4K | throughput vs F-Score: 0.5850 -> 0.6426; 54.0900 -> 53.7528 | Mixed |
Closest papers by meaning
- CacheQuant: Comprehensively Accelerated Diffusion Models (CVPR 2025 · similarity 0.89)
- DeepCache: Accelerating Diffusion Models for Free (CVPR 2024 · similarity 0.89)
- Cache Me if You Can: Accelerating Diffusion Models through Block Caching (CVPR 2024 · similarity 0.88)
- ManifoldCache: Training-Free Diffusion Acceleration via Constraint Manifold Caching (NeurIPS 2026 · similarity 0.88)
- DreamCraft3D: Hierarchical 3D Generation with Bootstrapped Diffusion Prior (ICLR 2024 · similarity 0.87)
- Learning-to-Cache: Accelerating Diffusion Transformer via Layer Caching (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
