metric_utils

Metrics Utils API

This module provides utility functions to compute metrics for classification tasks.

Functions

Name	Description
compute_metrics	Compute set of metrics for y_true and y_pred
confusion_matrix	Compute confusion matrix using keras w/ addition to normalize

Machine-readable model

• loggerattribute
• compute_metricsfunctionCompute set of metrics for ytrue and ypred.
• confusion_matrixfunctionCompute confusion matrix using keras w/ addition to normalize.

attribute

logger

Python

helia_edge/metrics/metric_utils.py:18

logger = logging.getLogger(__name__)

function

compute_metrics

Python

Compute set of metrics for ytrue and ypred.

helia_edge/metrics/metric_utils.py:21

compute_metrics(metrics: list[keras.Metric], y_true: keras.KerasTensor, y_pred: keras.KerasTensor) -> dict[str, float]

Compute set of metrics for y_true and y_pred.

Example:

import helia_edge as helia

metrics = [keras.metrics.Accuracy('acc'), keras.metrics.Precision(0.5, name='precision')]
y_true = np.array([0, 1, 1, 0])
y_pred = np.array([0, 1, 0, 1])
results = helia.metrics.compute_metrics(metrics, y_true, y_pred)
print(results)

Parameters

Parameters of compute_metrics

Name	Type	Default	Description
metrics	list[keras.Metric]	Required	List of metrics
y_true	keras.KerasTensor	Required	True labels
y_pred	keras.KerasTensor	Required	Predicted labels

Returns

Returns of compute_metrics

Value	Type	Description
dict	dict[str, float]	Dictionary of metric names and values

function

confusion_matrix

Python

Compute confusion matrix using keras w/ addition to normalize.

helia_edge/metrics/metric_utils.py:55

confusion_matrix(
    labels: keras.KerasTensor,
    predictions: keras.KerasTensor,
    num_classes: int,
    weights: keras.KerasTensor | None = None,
    dtype: str = 'int32',
    normalize: Literal['true', 'pred', 'all'] | None = None,
) -> keras.KerasTensor

Compute confusion matrix using keras w/ addition to normalize.

Normalization modes

• “true”: Normalize by true labels
• “pred”: Normalize by predicted labels
• “all”: Normalize by all labels

Example:

import helia_edge as helia

labels = np.array([0, 1, 1, 0])
predictions = np.array([0, 1, 0, 1])
num_classes = 2
cm = helia.metrics.confusion_matrix(labels, predictions, num_classes, normalize='true')
print(cm)

Parameters

Parameters of confusion_matrix

Name	Type	Default	Description
labels	keras.KerasTensor	Required	True labels
predictions	keras.KerasTensor	Required	Predicted labels
num_classes	int	Required	Number of classes
weights	keras.KerasTensor	None	Weights. Defaults to None.
dtype	str	'int32'	Data type. Defaults to "int32".
normalize	Literal['true', 'pred', 'all']	None	Normalization mode. Defaults to None.

Returns

Returns of confusion_matrix

Type	Description
keras.KerasTensor	keras.KerasTensor: Confusion matrix