Wav2Vec2 Classifier

speechline.classifiers.wav2vec2.Wav2Vec2Classifier (AudioClassifier)

Audio classifier with feature extractor.

Parameters:

Name	Type	Description	Default
model_checkpoint	str	HuggingFace model hub checkpoint.	required
max_duration_s	float	Maximum audio duration in seconds.	required

Source code in speechline/classifiers/wav2vec2.py

class Wav2Vec2Classifier(AudioClassifier):
    """
    Audio classifier with feature extractor.

    Args:
        model_checkpoint (str):
            HuggingFace model hub checkpoint.
        max_duration_s (float):
            Maximum audio duration in seconds.
    """

    def __init__(self, model_checkpoint: str, max_duration_s: float) -> None:
        super().__init__(model_checkpoint, max_duration_s=max_duration_s)

    def predict(self, dataset: Dataset) -> List[str]:
        """
        Performs audio classification (inference) on `dataset`.
        Preprocesses datasets, performs inference, then returns predictions.

        Args:
            dataset (Dataset):
                Dataset to be inferred.

        Returns:
            List[str]:
                List of predictions (as strings of labels).
        """
        return self.inference(dataset)

predict(self, dataset)

Performs audio classification (inference) on dataset. Preprocesses datasets, performs inference, then returns predictions.

Parameters:

Name	Type	Description	Default
dataset	Dataset	Dataset to be inferred.	required

Returns:

Type	Description
List[str]	List of predictions (as strings of labels).

Source code in speechline/classifiers/wav2vec2.py

def predict(self, dataset: Dataset) -> List[str]:
    """
    Performs audio classification (inference) on `dataset`.
    Preprocesses datasets, performs inference, then returns predictions.

    Args:
        dataset (Dataset):
            Dataset to be inferred.

    Returns:
        List[str]:
            List of predictions (as strings of labels).
    """
    return self.inference(dataset)