VastAI.create_workergroup

Create a new autoscale worker group.

Signature

VastAI.create_workergroup(
    template_hash: Optional[str] = None,
    template_id: Optional[int] = None,
    no_default: bool = False,
    launch_args: Optional[str] = None,
    endpoint_name: Optional[str] = None,
    endpoint_id: Optional[int] = None,
    test_workers: int = 3,
    gpu_ram: Optional[float] = None,
    search_params: Optional[str] = None,
    min_load: Optional[float] = None,
    target_util: Optional[float] = None,
    cold_mult: Optional[float] = None,
    cold_workers: Optional[int] = None
) -> str

Parameters

template_hash · Optional[str]

Hash of the instance template to use for workers.

template_id · Optional[int]

ID of the instance template to use for workers.

no_default · bool · default: "False"

Disable the default search query filters when finding offers.

launch_args · Optional[str]

Additional CLI arguments passed when launching worker instances.

endpoint_name · Optional[str]

Human-readable name for the endpoint this group belongs to.

endpoint_id · Optional[int]

ID of the endpoint this worker group belongs to.

test_workers · int · default: "3"

Number of test workers to spin up during initial scaling evaluation.

gpu_ram · Optional[float]

Minimum GPU RAM in GiB required per worker.

search_params · Optional[str]

Query string for filtering available GPU offers (e.g. “gpu_name=RTX_4090 num_gpus=1”).

min_load · Optional[float]

Minimum load (requests/sec) below which idle workers are scaled down.

target_util · Optional[float]

Target worker utilization ratio (0.0-1.0); workers are added to stay at this level.

cold_mult · Optional[float]

Multiplier applied to warm worker count for pre-warming cold workers.

cold_workers · Optional[int]

Number of pre-warmed (cold) workers to keep ready.

Returns

str, Result from the API call.

Example

from vastai import VastAI

client = VastAI(api_key="YOUR_API_KEY")
result = client.create_workergroup()
print(result)