Run evaluations on large datasets in Playgrounds or the SDK

import { Eval, initDataset, invoke } from "braintrust";

const project = "example-project";

await Eval(project, {
  experimentName: "large-run",
  data: initDataset(project, { dataset: "my-dataset" }),
  task: async (input) =>
    invoke({
      projectName: project,
      slug: "my-prompt",
      input,
    }),
  scores: [],
  maxConcurrency: 4,
});

from braintrust import Eval, init_dataset, invoke

project = "example-project"

Eval(
    project,
    experiment_name="large-run",
    data=init_dataset(project=project, name="my-dataset"),
    task=lambda input: invoke(
        project_name=project,
        slug="my-prompt",
        input=input,
    ),
    scores=[],
    max_concurrency=4,
)