Non-Parametric Quantization

I wrote a blog about the motivation behind this series of work.

Λ24-SQ-ViT

Spherical Leech Quantization for Visual Tokenization and Generation
Yue Zhao^1,2, Hanwen Jiang^1,3, Zhenlin Xu⁴, Chutong Yang¹, Ehsan Adeli², Philipp Krähenbühl^1
1UT Austin, ²Stanford University, ³Adobe Research, ⁴Mistral AI
arxiv | bibtex | Image Generation with Infinity+L24SQ

🍰 BSQ-ViT

You can pronouce BSQ-ViT like "biskvit" (a kind of Russian sponge cake) or simply "biscuit".

Image and Video Tokenization with Binary Spherical Quantization
Yue Zhao¹, Yuanjun Xiong², Philipp Krähenbühl^1
1UT Austin, ²Predera
arxiv | bibtex

Installation

1. Install Miniforge3

2. Create the environment

mamba env create -f bsqvit-env.yaml
mamba activate bsqvit

Main Results

Image Reconstruction (IN-1K val 256x256)

Video Reconstruction (UCF-101 16x128x128)

Image Synthesis (IN-1K 128x128)

Video Compression (MCL-JCV 360P)

License

MIT License.

Citing NPQ-ViT

@article{zhao2024lsqvit,
  title={Spherical Leech Quantization for Visual Tokenization and Generation},
  author={Zhao, Yue and Jiang, Hanwen and Xu, Zhenlin and Yang, Chutong and Adeli, Ehsan and Kr{\"a}henb{\"u}hl, Philipp},
  journal={arXiv preprint arXiv:2512.14697},
  year={2025}
}

@article{zhao2024bsqvit,
  title={Image and Video Tokenization with Binary Spherical Quantization},
  author={Zhao, Yue and Xiong, Yuanjun, and Kr{\"a}henb{\"u}hl, Philipp},
  journal={arXiv preprint arXiv:2406.07548},
  year={2024}
}