CoreML runs FP8-weight model on M6 chip in early patch test

AIFlow_ML · x · 2026-09-24

Developer @anemll shipped a patched CoreML model that loads FP8 weights on Apple's M6 chip, verified working on small tests. Larger model conversion is still pending, marking an early experiment in on-device FP8 deployment.

Original post →

More from Infra

Infra channel →