mlx-bun library API
Documentation
Preparing search index...
@mlx-bun/inference
@mlx-bun/inference/execution/fit
TRANSIENT_PER_TOKEN
Variable TRANSIENT_PER_TOKEN
Const
TRANSIENT_PER_TOKEN
:
550000
= 0.55e6
Prefill transient bytes per chunk token (measured: ~1.1 GB @ 2048).
Settings
Member Visibility
Protected
Inherited
External
Theme
OS
Light
Dark
Documentation
mlx-bun library API
Loading...
Prefill transient bytes per chunk token (measured: ~1.1 GB @ 2048).