mlx-omarchy v0.5.0 released with a claimed first end-to-end Core ML model run on Linux
The release announcement says Parakeet matches the macOS transcript byte for byte for on-device voice, with its encoder running on the Apple Neural Engine.
TLDR
The September 15 announcement of mlx-omarchy v0.5.0 claims the first end-to-end run of a Core ML model on Linux. It says Parakeet’s encoder runs on the Apple Neural Engine, with its on-device voice transcript matching macOS byte for byte. The announcement lists M1 and M1 Max as fully supported, with M2+ support still a work in progress.
Combined views
9K
1 Source, first seen 15d ago