Open Models Almanac
Methodsinference & servinginference kernel

single mention · inference & serving

Persistent FlyDSL fused MoE decode kernel

Runs MoE layers for a decode step as a persistent FlyDSL kernel.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See inference kernel for the methods filed beside it.

Also called persistent FlyDSL kernel.

Filed under inference & serving › inference kernel