Wavelet Latent Diffusion (WaLa): Billion-Parameter 3D Generative Model with Compact Wavelet Encodings

ICLR 2025 Conference Submission480 Authors

Keywords: 3D generative modelling, Diffusion models, wavelet encoding

TL;DR: We introduce WaLa, a method that tackles the dimensional and computational challenges of 3D generation with impressive compression while maximizing fidelity.

Abstract: Large-scale 3D generative models require substantial computational resources yet often fall short in capturing fine details and complex geometries at high resolutions. We attribute this limitation to the inefficiency of current representations, which lack the compactness required for generative networks to model effectively. To address this, we introduce Wavelet Latent Diffusion (WaLa), a novel approach that encodes 3D shapes into a wavelet-based, compact latent encodings. Specifically, we compress a $256^3$ signed distance field into a $12^3 \times 4$ latent grid, achieving an impressive 2,427× compression ratio with minimal loss of detail. This high level of compression allows our method to efficiently train large-scale generative networks without increasing inference time. Our models, both conditional and unconditional, contain approximately one billion parameters and successfully generate high-quality 3D shapes at $256^3$ resolution. Moreover, WaLa offers rapid inference, producing shapes within 2–4 seconds depending on the condition, despite the model’s scale. We demonstrate state-of-the-art performance across multiple datasets, with significant improvements in generation quality, diversity, and computational efficiency. Upon acceptance, we will open-source the code and model weights for public use and reproducibility.

Primary Area: generative models

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 480