TimeSeriesFoundationModel

class autogluon.cloud.TimeSeriesFoundationModel(model_id: str, **kwargs)

Pretrained time series foundation model for zero-shot forecasting on Amazon SageMaker.

Wraps pretrained models like Chronos-2, Chronos-Bolt, and Toto-2.0, with no training required. Use list_models() to get the supported model_id values, and see the foundation model tutorial for a full walkthrough.

Predictions can be produced in three modes:

• Batch — predict() runs a one-off SageMaker training job and writes forecasts to S3.

• Real-time — deploy() provisions a real-time endpoint; call TimeSeriesEndpoint.predict() for low-latency inference, then TimeSeriesEndpoint.delete_endpoint() to tear it down.

• Serverless — deploy() with inference_mode="serverless" provisions a SageMaker Serverless Inference endpoint that scales to zero. Requires a cached model artifact (see cache_model_artifact()).

Parameters:

• model_id (str) – ID of the foundation model, e.g. "chronos-2" or "mitra-classifier". Use list_models() to get the supported IDs.

• cloud_output_path (str | None, default = None) –

  S3 location where intermediate artifacts are stored. Accepts:

  • s3://bucket — a unique timestamped subfolder ag-<timestamp> is appended.

  • s3://bucket/prefix — used verbatim. Re-running with the same prefix will overwrite previously written artifacts.

  • None (default) — use the bucket saved in ~/.autogluon/cloud.yaml (set by autogluon.cloud.bootstrap() / autogluon.cloud.register()) and append a timestamped subfolder. Raises if no bucket is configured.

• role (str | None, default = None) – ARN of the SageMaker execution role used to run training and inference jobs. If None, falls back to role_arn in ~/.autogluon/cloud.yaml (set by autogluon.cloud.bootstrap() / autogluon.cloud.register()), and finally to the role of the current AWS identity.

• hyperparameters (dict[str, Any] | None, default = None) – Model hyperparameters. Hyperparameters passed to predict() or deploy() take precedence.

• model_artifact_uri (str | None, default = None) – S3 URI of a model.tar.gz created by cache_model_artifact(). If set, deploy() loads the weights from this artifact instead of downloading them from Hugging Face.

• backend (Literal["sagemaker"], default = "sagemaker") – Cloud backend to use.

Methods

cache_model_artifact	Bundle the model weights from Hugging Face and the serve script into a model.tar.gz on S3.
deploy	Deploy the model to a real-time or serverless endpoint.
from_dict	Create a model from the output of to_dict().
from_json	Create a model from the output of to_json().
list_models	List the supported model_id values.
predict	Forecast the future values of data in a one-off SageMaker job.
to_dict	Serialize model_id, hyperparameters, and model_artifact_uri to a dict.
to_json	Serialize model_id, hyperparameters, and model_artifact_uri to a JSON string.