Document why MLX is not shipped: 2.9x slower than Core ML and non-finite output

9a16f90

verified

msluszniak commited on 11 days ago