components compute_metrics

Compute Metrics

compute_metrics

Overview

Calculate model performance metrics, given ground truth and prediction data.

Version: 0.0.36

Tags

type : evaluation sub_type : compute_metrics

View in Studio: https://ml.azure.com/registries/azureml/components/compute_metrics/version/0.0.36

Inputs

Name	Description	Type	Default	Optional	Enum
task	Task type	string	tabular-classification	False	['tabular-classification', 'tabular-classification-multilabel', 'tabular-regression', 'tabular-forecasting', 'text-classification', 'text-classification-multilabel', 'text-named-entity-recognition', 'text-summarization', 'question-answering', 'text-translation', 'text-generation', 'fill-mask', 'image-classification', 'image-classification-multilabel', 'chat-completion', 'image-object-detection', 'image-instance-segmentation']
ground_truth	Ground Truths of Test Data as a 1-column JSON Lines file	uri_folder		True
ground_truth_column_name	Column name which contains ground truths in provided uri file for ground_truth. (Optional if we have one column name.)	string		True
prediction	Model Predictions as a 1-column JSON Lines file	uri_folder		False
prediction_column_name	Column name which contains ground truths in provided uri file for prediction. (Optional if we have one column name.)	string		True
prediction_probabilities	Predictions Probabilities as 1-column JSON Lines file	uri_folder		True
evaluation_config	Additional parameters required for evaluation.	uri_file		True
evaluation_config_params	JSON Serialized string of evaluation_config	string		True
openai_config_params	Required OpenAI Params for calculating GPT Based metrics for QnA task	string		True

Outputs

Name	Description	Type
evaluation_result		uri_folder

Environment

azureml://registries/azureml/environments/model-evaluation/labels/latest