forked from third-party-mirrors/ollama
3 lines
372 B
Markdown
3 lines
372 B
Markdown
# `discover`
|
|
|
|
This package is responsible for discovering information about the system and the capabilities to run LLM. This includes GPU and CPU discovery so the optimal runner can be chosen for a given model. The ollama scheduler relies on up-to-date available memory information, so this package provides the ability to refresh free memory as efficiently as possible. |