Skip to content

AI Developer Tool · Python

gauss

GAUSS: Distribution-Aware Compression for Neural Network Weights

What it is

Gauss is a tool for compressing LLM checkpoints using Gaussian Mixture Model and Asymmetric Numeral Systems entropy coding, achieving ~4.9x compression with bounded error. It provides a CLI and Python API for lossless storage of small tensors and random access to individual tensors.

  • compression-algorithm
  • deep-learning
  • llm
  • ml
  • neural-network
Scored at143295518 Nov 2026Changes since

Merit describes the code at this commit. Later pushes count once the owner rescans. How merit is scored

Distribution based on AI work

1 / 2
Machine Learning: 45.9%NLP & LLMs: 35.1%MLOps: 16.2%Computer Vision: 2.7%
  • Machine Learning46%
  • NLP & LLMs35%
  • MLOps16%
  • Computer Vision3%

Each field's share of the project's AI work, as scoring counts it. It ranks in the field it is most about.

What the judge found

Read from the project's code at its scored commit.

Strengths4

  • High compression ratio (~4.9x) for LLM checkpoints with bounded error.
  • CPU-only operation with parallel processing via multiprocessing.
  • Preserves original data types exactly and supports random access to individual tensors.
  • Well-documented README with clear usage examples and explanations of the algorithm.

Limitations

The owner keeps this project's limitations private.

Around it on the Machine Learning board

Whole board
  1. 23CreatorOS-StudioML Application · TypeScript63.0/100AdvancedthePloggy
  2. 24model-anatomyAI Developer Tool · Python62.9/100AdvancedPranav
  3. 25, shared rankmamba-viewAI Developer Tool · Python62.2/100AdvancedRick
  4. 25, shared rankResNet_Family-from-scratchArchitecture Reimplementation · Jupyter Notebook62.2/100AdvancedPop Alexandru
  5. 25, shared rankgaussAI Developer Tool · Python62.2/100AdvancedMinsung Skystarry

More from Minsung Skystarry