⬢github Python · 20 ★ +14 since we first saw it · pushed 1 d ago · NOASSERTION
NeuraCryption/SPECTRA
Compress any LLM in seconds, load it for inference and in one click you have an AI that could have taken hours of vLLM configuration and what not to setup and load. Of course there are opensource compressed models but this is for people who have build their own AI, finetuned it, etc.
SPECTRA is a Python tool that compresses LLMs (transformers, MoE, Mamba hybrids, conv nets) to 4-bit with minimal accuracy loss, then serves them on low-VRAM NVIDIA GPUs like 4 GB laptop cards. It promises fast compression without calibration data, near-instant model loading, and one-click inference with Triton-based GPU kernels.
Why now: It was just shared on Hacker News with a bold claim: 3x compression, near-lossless quality, under 50 MB RAM during compression — likely drawing curiosity and skepticism.
Who it is for: Hobbyists and researchers with custom or fine-tuned models who want to run them locally on small consumer GPUs.
Stars over our 229 snapshots: 6 to 20, since 2 d ago.
Where people talked about it
API: https://socialmediatrends-api.osmike.com/v1/repos/NeuraCryption/SPECTRA