Synapse Open Synapse
Tech

Compress any LLM 3x in one click,near-lossless,in under 50 MB RAM

HN · · 38 min read

The peer-reviewed publication of SPECTRA (Zenodo DOI: 10.5281/zenodo.24058202). 3.31x model compression, calibration-free Group-32 microscaling, and 82.53 tok/s decoding on 4GB consumer GPUs.

Latest from HN