espnet.nets.pytorch_backend.frontends.feature_transform.FeatureTransform

source

class espnet.nets.pytorch_backend.frontends.feature_transform.FeatureTransform(fs: int = 16000, n_fft: int = 512, n_mels: int = 80, fmin: float = 0.0, fmax: float | None = None, stats_file: str | None = None, apply_uttmvn: bool = True, uttmvn_norm_means: bool = True, uttmvn_norm_vars: bool = False)

Bases: Module

Feature transform class.

Initialize Feature transform.

forward(x: ComplexTensor, ilens: LongTensor | ndarray | List[int]) → Tuple[Tensor, LongTensor]

Calculate feature transform forward propagation.