thiagosilvahyper/bihe-quantization

BIHE Protocol - Next-generation vector quantization combining E8 lattice geometry with Lloyd algorithm optimization. Achieves 16× compression ratio while maintaining 88.5% recall on real-world datasets. Validated on Stanford SQuAD v2.0 (5,000 texts) and Microsoft MS MARCO (6,000 texts). Features NSM below Shannon theoretical limits.

17
/ 100
Experimental
No Package No Dependents
Maintenance 6 / 25
Adoption 2 / 25
Maturity 9 / 25
Community 0 / 25

How are scores calculated?

Stars

2

Forks

—

Language

Python

License

MIT

Category

self-hosted-embedding-servers

Last pushed

Oct 27, 2025

Commits (30d)

0

GitHub
Self-Hosted Embedding Servers · 83 tools

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/thiagosilvahyper/bihe-quantization"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.

Featured in

Embeddings Are Easier Than Whatever You're Doing Instead

Higher-rated alternatives

FlagOpen/FlagEmbedding

Retrieval and Retrieval-augmented LLMs

79

Blaizzy/mlx-embeddings

MLX-Embeddings is the best package for running Vision and Language Embedding models locally on...

76

qdrant/fastembed

Fast, Accurate, Lightweight Python library to make State of the Art Embedding

74

Merck/Sapiens

Sapiens is a human antibody language model based on BERT.

69

amansrivastava17/embedding-as-service

One-Stop Solution to encode sentence to fixed length vectors from various embedding techniques

53

Explore Embedding Tools

All categories Trending Embeddings directory Insights