Open Models Almanac
Methodsoptimizationtraining runtime

single mention · optimization

Seeding node-local storage from a warm shared cache

Extracts warm shared-cache artifacts into node-local storage before Ray initialization.

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

Also called Node-local seeding at startup.

Filed under optimization › training runtime