I want to run batch inference on sagemaker orchestrated by Airflow using sagemakertransform operator using an existing model from model registry. For the operator it takes model name as input.