Skip to content

Quantization Lab

TaxCollector23 edited this page Oct 10, 2026 · 2 revisions

Quantization Lab — not shipped

AXIOM does not currently ship a quantization runner, inference bundle, or throughput target. Its current planning commands can describe a conservative QLoRA or quantization-adjacent starting point, but they do not transform model weights or claim tokens-per-second results.

Use Training for the supported planning path. Keep any future quantization workflow separate from planning until its inputs, outputs, and measurement method are implemented and tested.

Clone this wiki locally