components batch_output_formatter

Batch Output Formatter

batch_output_formatter

Overview

Output Formatter for batch inference output

Version: 0.0.12

View in Studio: https://ml.azure.com/registries/azureml/components/batch_output_formatter/version/0.0.12

Inputs

Name	Description	Type	Default	Optional	Enum
model_type	Type of model. Can be one of ('oai', 'oss', 'vision_oss', 'claude')	string		True
batch_inference_output	The raw batch inference output.	uri_folder		False
label_column_name	The label column name.	string		True
additional_columns	Name(s) of additional column(s) that could be useful to compute metrics, separated by comma (",").	string		True
endpoint_url		string		True
ground_truth_input	The raw batch inference output.	uri_folder		True
handle_response_failure	The way that the formatter handles the failed response.	string	use_fallback	False	['use_fallback', 'neglect']
fallback_value	The fallback value that can be used when request payload failed.	string		True
min_endpoint_success_ratio	The minimum value of (successful_requests / total_requests) required for classifying inference as successful. If (successful_requests / total_requests) < min_endpoint_success_ratio, the experiment will be marked as failed. By default it is 0. (0 means all requests are allowed to fail while 1 means no request should fail.)	number	0	False
is_performance_test	If true, the performance test will be run.	boolean	False	False
use_tiktoken	If true, cl100k_base encoder is used from tiktoken to calculate token count; overrides any other token count calculation.	boolean	False	True

Outputs

Name	Description	Type
predictions		uri_file
performance_metadata		uri_file
ground_truth		uri_file
successful_requests		uri_file
failed_requests		uri_file
unsafe_content_blocked_requests		uri_file

Environment

azureml://registries/azureml/environments/model-evaluation/versions/34