hezar.metrics.rouge module

class hezar.metrics.rouge.ROUGE(config: ROUGEConfig, **kwargs)

Bases: Metric

ROUGE metric for evaluating text summarization using rouge_score.

Parameters:

• config (ROUGEConfig) – Metric configuration object.

• **kwargs – Extra configuration parameters passed as kwargs to update the config.

compute(predictions=None, targets=None, use_aggregator=None, n_decimals=None, output_keys=None, **kwargs)

Computes the ROUGE scores for the given predictions against targets.

Parameters:

• predictions – Predicted summaries.

• targets – Ground truth summaries.

• use_aggregator (bool) – Flag to enable score aggregation for multiple references.

• n_decimals (int) – Number of decimals for the final score.

• output_keys (tuple) – Filter the output keys.

Returns:

A dictionary of the metric results, with keys specified by output_keys.

Return type:

dict

required_backends: List[str | Backends] = [Backends.ROUGE]

class hezar.metrics.rouge.ROUGEConfig(objective: str = 'maximize', output_keys: tuple = ('rouge1', 'rouge2', 'rougeL', 'rougeLsum'), n_decimals: int = 4, use_stemmer: bool = False, use_aggregator: bool = True, multi_ref: bool = True)

Bases: MetricConfig

Configuration class for ROUGE metric.

Parameters:

• name (MetricType) – The type of metric, ROUGE in this case.

• use_stemmer (bool) – Flag to enable stemming when computing ROUGE.

• use_aggregator (bool) – Flag to enable score aggregation for multiple references.

• multi_ref (bool) – Flag to indicate if multiple references are present.

• output_keys (tuple) – Keys to filter the metric results for output.

multi_ref: bool = True

name: str = 'rouge'

objective: str = 'maximize'

output_keys: tuple = ('rouge1', 'rouge2', 'rougeL', 'rougeLsum')

use_aggregator: bool = True

use_stemmer: bool = False