components batch_deploy_model

Batch deploy model

batch_deploy_model

Overview

Batch deploy a model to a workspace. The component works on compute with MSI attached.

Version: 0.0.5

Tags

Preview Internal

View in Studio: https://ml.azure.com/registries/azureml/components/batch_deploy_model/version/0.0.5

Inputs

Output of registering component

Name	Description	Type	Default	Optional	Enum
registration_details_folder	Folder containing model registration details in a JSON file named model_registration_details.json	uri_folder		True
model_id	Asset ID of the model registered in workspace/registry. Registry - azureml://registries//models//versions/ Workspace - azureml::	string		True
inference_payload_file	File containing data used to validate deployment	uri_file		True
inference_payload_folder	Folder containing files used to validate deployment	uri_folder		True
endpoint_name	Name of the endpoint	string		True
deployment_name	Name of the deployment	string	default	True
compute_name	Name of the compute cluster to execute the batch scoring jobs on. New compute will be created if the compute cluster is not present.	string	cpu-cluster	True
size	Compute instance size to deploy model. Make sure that instance type is available and have enough quota available.	string	Standard_NC24s_v3	True	['Standard_DS1_v2', 'Standard_DS2_v2', 'Standard_DS3_v2', 'Standard_DS4_v2', 'Standard_DS5_v2', 'Standard_F2s_v2', 'Standard_F4s_v2', 'Standard_F8s_v2', 'Standard_F16s_v2', 'Standard_F32s_v2', 'Standard_F48s_v2', 'Standard_F64s_v2', 'Standard_F72s_v2', 'Standard_FX24mds', 'Standard_FX36mds', 'Standard_FX48mds', 'Standard_E2s_v3', 'Standard_E4s_v3', 'Standard_E8s_v3', 'Standard_E16s_v3', 'Standard_E32s_v3', 'Standard_E48s_v3', 'Standard_E64s_v3', 'Standard_NC4as_T4_v3', 'Standard_NC6s_v2', 'Standard_NC6s_v3', 'Standard_NC8as_T4_v3', 'Standard_NC12s_v2', 'Standard_NC12s_v3', 'Standard_NC16as_T4_v3', 'Standard_NC24s_v2', 'Standard_NC24s_v3', 'Standard_NC24rs_v3', 'Standard_NC64as_T4_v3', 'Standard_ND40rs_v2', 'Standard_ND96asr_v4', 'Standard_ND96amsr_A100_v4']
min_instances	Minimum number of instances of the compute cluster to be created.	integer	0	True
max_instances	Maximum number of instances of the compute cluster to be created.	integer	1	True
idle_time_before_scale_down	Node Idle Time before scaling down the compute cluster to be created.	integer	120	True
output_file_name	Name of the batch scoring output file.	string	predictions.csv	True
max_concurrency_per_instance	The maximum number of parallel scoring_script runs per instance.	integer	1	True
error_threshold	The number of file failures that should be ignored.	integer	-1	True
max_retries	The maximum number of retries for a failed or timed-out mini batch.	integer	3	True
timeout	The timeout in seconds for scoring a single mini batch.	integer	500	True
logging_level	The log verbosity level.	string	info	True
mini_batch_size	The number of files the code_configuration.scoring_script can process in one run() call.	integer	10	True
instance_count	The number of nodes to use for each batch scoring job.	integer	1	True

Outputs

Name	Description	Type
batch_job_output_folder	Folder to which batch job outputs will be saved.	uri_folder

Environment

azureml://registries/azureml/environments/python-sdk-v2/versions/19