mlx-bun library API
    Preparing search index...

    Interface BenchmarkTask

    One task the runner can evaluate (eval-serve.ts tasks).

    interface BenchmarkTask {
        component: string | null;
        datasets: readonly {
            name: string;
            present: boolean;
            rows: number;
            sha256: string;
            source: string;
        }[];
        id: string;
        needsVerifier: boolean;
        sets: readonly string[];
    }
    Index
    component datasets id needsVerifier sets
    component: string | null

    The capability component the task feeds; the smoketest subset has none.

    datasets: readonly {
        name: string;
        present: boolean;
        rows: number;
        sha256: string;
        source: string;
    }[]

    The pinned dataset files the task reads, and whether each is in the module's data directory.

    id: string
    needsVerifier: boolean

    Executes generated code, so it needs the Docker verifier; without one it is skipped as unverified.

    sets: readonly string[]

    The task sets that contain it: capability, smoketest, all.

    Settings
    Member Visibility

    On This Page
    Properties