TabularCloudPredictor

class autogluon.cloud.TabularCloudPredictor(local_output_path: str | None = None, cloud_output_path: str | None = None, backend: str = 'sagemaker', role: str | None = None, verbosity: int = 2)

Train and deploy AutoGluon tabular models (classification and regression) on AWS SageMaker.

Wraps autogluon.tabular.TabularPredictor (docs) and runs fit, predict, and endpoint deployment as managed SageMaker jobs.

Parameters:

• local_output_path (Optional[str], default = None) – Path to directory where downloaded trained predictor, batch transform results, and intermediate outputs should be saved If unspecified, a time-stamped folder called “AutogluonCloudPredictor/ag-[TIMESTAMP]” will be created in the working directory to store all downloaded trained predictor, batch transform results, and intermediate outputs. Note: To call fit() twice and save all results of each fit, you must specify different local_output_path locations or don’t specify local_output_path at all. Otherwise files from first fit() will be overwritten by second fit().

• cloud_output_path (Optional[str], default = None) –

  S3 location where intermediate artifacts and trained models are stored. Accepts:

  • s3://bucket — a unique timestamped subfolder ag-<timestamp> is appended, so each call gets its own folder and repeated runs don’t overwrite each other.

  • 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.

• backend (str, default = "sagemaker") – The backend to use. Valid options are: “sagemaker” and “ray_aws”. SageMaker backend supports training, deploying and batch inference on AWS SageMaker. Only single instance training is supported. RayAWS backend supports distributed training by creating an ephemeral ray cluster on AWS. Deployment and batch inferenc are not supported yet.

• role (Optional[str], 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 sagemaker.get_execution_role().

• verbosity (int, default = 2) – Verbosity levels range from 0 to 4 and control how much information is printed. Higher levels correspond to more detailed print statements (you can set verbosity = 0 to suppress warnings). If using logging, you can alternatively control amount of information printed via logger.setLevel(L), where L ranges from 0 to 50 (Note: higher values of L correspond to fewer print statements, opposite of verbosity levels).

Methods

attach_endpoint	Attach the current CloudPredictor to an existing endpoint.
attach_job	Attach to a sagemaker training job.
cleanup_deployment	Delete the deployed endpoint and other artifacts
deploy	Deploy a predictor to an endpoint, which can be used to do real-time inference later.
detach_endpoint	Detach the current endpoint and return it.
download_trained_predictor	Download the trained predictor from the cloud.
fit	Fit the predictor with the backend.
get_batch_inference_job_info	Get general info of the batch inference job.
get_batch_inference_job_status	Get the status of the batch inference job.
get_fit_job_output_path	Get the output path in the cloud of the trained artifact
get_fit_job_status	Get the status of the training job.
info	Return general info about CloudPredictor
leaderboard
load	Load the CloudPredictor
predict	Batch inference.
predict_proba	Batch inference When minimizing latency isn't a concern, then the batch transform functionality may be easier, more scalable, and more appropriate.
predict_proba_real_time	Predict probability with the deployed endpoint.
predict_real_time	Predict with the deployed endpoint.
save	Save the CloudPredictor so that user can later reload the predictor to gain access to deployed endpoint.
to_local_predictor	Convert the Cloud trained predictor to a local AutoGluon Predictor.

Attributes

backend_map
endpoint_name	Return the CloudPredictor deployed endpoint name
is_fit	Whether this CloudPredictor is fitted already
predictor_file_name
predictor_type	Type of the underneath AutoGluon Predictor