maple.utils.eval.EvalResult

class maple.utils.eval.EvalResult(run_id: str, policy_id: str, env_id: str, task: str, instruction: str, seed: int, steps: int = 0, total_reward: float = 0.0, success: bool = False, terminated: bool = False, truncated: bool = False, started_at: float = 0.0, finished_at: float = 0.0, duration_seconds: float = 0.0, video_path: str | None = None, error: str | None = None, metadata: Dict[str, ~typing.Any]=<factory>)

Result container for a single evaluation episode.

Stores all information about an episode execution including outcomes, timing, and optional error information. Provides serialization for persistence and reporting.

Methods

__init__(run_id, policy_id, env_id, task, ...)

from_dict(d)

Create result from dictionary representation.

to_dict()

Convert result to dictionary representation.

Attributes

duration_seconds

error

finished_at

started_at

steps

success

terminated

total_reward

truncated

video_path

run_id

policy_id

env_id

task

instruction

seed

metadata