Skip to main content

run_eval

Function run_eval 

Source
pub async fn run_eval(
    options: &EvalOptions,
    tasks: &[MineTask],
    writer: &mut JsonlWriter,
) -> Result<EvalSummary>
Expand description

Stream every task to the runner and write one attempt per reply.

Failures are converted into counts and log lines rather than aborting the run, except when the runner cannot be started at all.