llama.backend

Undocumented in source.

Members

Functions

initBackend
BackendGuard initBackend()

Initialises the llama backend and returns a guard that frees it on scope exit.

loadAllBackends
void loadAllBackends()

Loads all available ggml backends (CPU, CUDA, Metal, etc.).

Structs

BackendGuard
struct BackendGuard

Calls llama_backend_free when it goes out of scope.

Meta