Daemoniorum-LLC/haagenti
High-performance compression library for Rust with LZ4, Zstd, Brotli, and Deflate. Features SIMD acceleration, streaming API, and no_std support.
This project helps AI practitioners run very large language models (like 70 billion parameter models) on a single consumer-grade graphics card, instead of requiring expensive, specialized cloud servers or multiple high-end GPUs. It takes your massive AI model weights and transforms them into a highly compressed format, allowing them to fit within limited GPU memory. This is for AI engineers, researchers, or data scientists who need to deploy or experiment with large AI models more affordably.
Use this if you need to perform inference with large AI models (e.g., 70B+ parameters) on hardware with limited VRAM, such as a single consumer GPU, while maintaining output quality.
Not ideal if your primary need is general-purpose data compression for typical files or network traffic, where traditional algorithms are sufficient.
Stars
23
Forks
1
Language
Rust
License
Apache-2.0
Category
Last pushed
Mar 09, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ai-coding/Daemoniorum-LLC/haagenti"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.