RentResult

from lium.sdk import RentResult

Defined in lium.sdk.models.

What rent() chose and, unless it was a dry run, rented.

pod is the same dict up() returns (id, name, executor_id, …) and is None on a dry run. gpu_count is the GPUs rented and price_per_hour what the rental bills (price_per_gpu × gpu_count); both differ from executor.gpu_count / executor.price_per_hour when the server rents a split of a larger node. alternatives are the runners-up in the order they would have been tried; attempts is how many rents the server made before one succeeded (0 on a dry run, > 1 when the first pick was taken meanwhile); server_side says whether the backend chose the node or this client did.

RentResult(
    executor: lium.sdk.models.ExecutorInfo,
    price_per_hour: float,
    gpu_count: int = 1,
    pod: Optional[Dict] = None,
    template_id: Optional[str] = None,
    candidates: int = 1,
    alternatives: List[Dict] = <factory>,
    attempts: int = 0,
    dry_run: bool = False,
    server_side: bool = False
)