# GeoCache: Training-Free Acceleration of Multi-View Texture Diffusion via Geometric Delta Transport

arXiv: https://arxiv.org/abs/2608.13255 · HTML: https://mlsanity.com/paper/2608.13255/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 26 of its 31 references that are in the map.
- First posted on arXiv 13 Aug 2026.

## Builds on (most cited first, 26 of 26)

- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Denoising Diffusion Implicit Models](https://mlsanity.com/paper/2010.02502/) (2020)
- [Objaverse: A Universe of Annotated 3D Objects](https://mlsanity.com/paper/2212.08051/) (2022)
- [Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference](https://mlsanity.com/paper/2310.04378/) (2023)
- [Google Scanned Objects: A High-Quality Dataset of 3D Scanned Household Items](https://mlsanity.com/paper/2204.11918/) (2022)
- [UniPC: A Unified Predictor-Corrector Framework for Fast Sampling of Diffusion Models](https://mlsanity.com/paper/2302.04867/) (2023)
- [ABO: Dataset and Benchmarks for Real-World 3D Object Understanding](https://mlsanity.com/paper/2110.06199/) (2021)
- [TEXTure: Text-Guided Texturing of 3D Shapes](https://mlsanity.com/paper/2302.01721/) (2023)
- [Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model](https://mlsanity.com/paper/2411.19108/) (2024)
- [MVDiffusion: Enabling Holistic Multi-view Image Generation with Correspondence-Aware Diffusion](https://mlsanity.com/paper/2307.01097/) (2023)
- [Delta-DiT: A Training-Free Acceleration Method Tailored for Diffusion Transformers](https://mlsanity.com/paper/2406.01125/) (2024)
- [From Reusing to Forecasting: Accelerating Diffusion Models with TaylorSeers](https://mlsanity.com/paper/2503.06923/) (2025)
- [Paint3D: Paint Anything 3D with Lighting-Less Texture Diffusion Models](https://mlsanity.com/paper/2312.13913/) (2023)
- [FasterCache: Training-Free Video Diffusion Model Acceleration with High Quality](https://mlsanity.com/paper/2410.19355/) (2024)
- [MV-Adapter: Multi-view Consistent Image Generation Made Easy](https://mlsanity.com/paper/2412.03632/) (2024)
- [Text-Guided Texturing by Synchronized Multi-View Diffusion](https://mlsanity.com/paper/2311.12891/) (2023)
- [TEXGen: a Generative Diffusion Model for Mesh Textures](https://mlsanity.com/paper/2411.14740/) (2024)
- [MagCache: Fast Video Generation with Magnitude-Aware Cache](https://mlsanity.com/paper/2506.09045/) (2025)
- [Hash3D: Training-free Acceleration for 3D Generation](https://mlsanity.com/paper/2404.06091/) (2024)
- [A Survey on Cache Methods in Diffusion Models: Toward Efficient Multi-Modal Generation](https://mlsanity.com/paper/2510.19755/) (2025)
- [CAMEO: Correspondence-Attention Alignment for Multi-View Diffusion Models](https://mlsanity.com/paper/2512.03045/) (2025)
- [MVPainter: Accurate and Detailed 3D Texture Generation via Multi-View Diffusion with Geometric Control](https://mlsanity.com/paper/2505.12635/) (2025)
- [SeqTex: Generate Mesh Textures in Video Sequence](https://mlsanity.com/paper/2507.04285/) (2025)
- [CaliTex: Geometry-Calibrated Attention for View-Coherent 3D Texture Generation](https://mlsanity.com/paper/2511.21309/) (2025)
- [Fast3Dcache: Training-free 3D Geometry Synthesis Acceleration](https://mlsanity.com/paper/2511.22533/) (2025)
- [MD-ProjTex: Texturing 3D Shapes with Multi-Diffusion Projection](https://mlsanity.com/paper/2504.02762/) (2025)

## Abstract (arXiv)

Geometry-conditioned multi-view diffusion enables high-quality 3D texture generation, but its repeated per-view denoiser evaluations introduce substantial computational cost. Existing training-free accelerators primarily exploit temporal redundancy by reusing computation across denoising steps. In multi-view texturing, however, skipping a step also removes the cross-view interaction that continually aligns different observations of the same surface, leading to rapidly degraded consistency and fidelity. Our analysis identifies a complementary source of redundancy: although intermediate features remain view-specific, geometrically corresponding surface points exhibit transferable evolution in their predicted clean signals. Based on this observation, we introduce gc, a training-free plugin that evaluates a rotating subset of anchor views and transports their geometry-aligned per-step xz updates to the remaining views. Periodic full-view computation controls accumulated error, while sampler-consistent reconstruction preserves the denoising trajectory. gc requires neither retraining nor architectural modification and uses the position maps already available in geometry-conditioned texturing pipelines. Across Hunyuan3D-2.1, SyncMVD, and MVPainter, gc achieves a stronger speed--fidelity trade-off than temporal caches and step reduction at operating points above 2×. On Hunyuan3D-2.1, it delivers a 2.21× denoiser-loop speedup with an MV-LPIPS of 0.0293 and an MV-PSNR of 33.60 dB, providing the best fidelity among all tested methods above 2×. The same transferred configuration reaches the highest speedup and lowest FLOPs on SyncMVD, while gc achieves the lowest FLOPs and best fidelity among the accelerated methods on MVPainter. These results establish cross-view geometry as an effective acceleration axis for multi-view texture diffusion.