espnet2.asvspoof package

espnet2.asvspoof.espnet_model

class espnet2.asvspoof.espnet_model.ESPnetASVSpoofModel(frontend: Optional[espnet2.asr.frontend.abs_frontend.AbsFrontend], specaug: Optional[espnet2.asr.specaug.abs_specaug.AbsSpecAug], normalize: Optional[espnet2.layers.abs_normalize.AbsNormalize], encoder: espnet2.asr.encoder.abs_encoder.AbsEncoder, preencoder: Optional[espnet2.asr.preencoder.abs_preencoder.AbsPreEncoder], decoder: espnet2.asvspoof.decoder.abs_decoder.AbsDecoder, losses: Dict[str, espnet2.asvspoof.loss.abs_loss.AbsASVSpoofLoss])

Bases: espnet2.train.abs_espnet_model.AbsESPnetModel

ASV Spoofing model

A simple ASV Spoofing model

collect_feats(speech: torch.Tensor, speech_lengths: torch.Tensor, **kwargs) → Dict[str, torch.Tensor]

encode(speech: torch.Tensor, speech_lengths: torch.Tensor) → Tuple[torch.Tensor, torch.Tensor]

Frontend + Encoder

Parameters:

• speech – (Batch, Length, …)

• speech_lengths – (Batch,)

• bottleneck_feats – (Batch, Length, …): used for enh + diar

forward(speech: torch.Tensor, speech_lengths: torch.Tensor = None, label: torch.Tensor = None, **kwargs) → Tuple[torch.Tensor, Dict[str, torch.Tensor], torch.Tensor]

Frontend + Encoder + Decoder + Calc loss

Parameters:

• speech – (Batch, samples)

• spk_labels – (Batch, )

• kwargs – “utt_id” is among the input.

espnet2.asvspoof.__init__

espnet2.asvspoof.decoder.abs_decoder

class espnet2.asvspoof.decoder.abs_decoder.AbsDecoder(*args, **kwargs)

Bases: torch.nn.modules.module.Module, abc.ABC

Initializes internal Module state, shared by both nn.Module and ScriptModule.

abstract forward(input: torch.Tensor, ilens: torch.Tensor) → Tuple[torch.Tensor, torch.Tensor]

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

espnet2.asvspoof.decoder.linear_decoder

class espnet2.asvspoof.decoder.linear_decoder.LinearDecoder(encoder_output_size: int)

Bases: espnet2.asvspoof.decoder.abs_decoder.AbsDecoder

Linear decoder for speaker diarization

forward(input: torch.Tensor, ilens: Optional[torch.Tensor])

Forward.

Parameters:

• input (torch.Tensor) – hidden_space [Batch, T, F]

• ilens (torch.Tensor) – input lengths [Batch]

espnet2.asvspoof.decoder.__init__

espnet2.asvspoof.loss.abs_loss

class espnet2.asvspoof.loss.abs_loss.AbsASVSpoofLoss(*args, **kwargs)

Bases: torch.nn.modules.module.Module, abc.ABC

Base class for all ASV Spoofing loss modules.

Initializes internal Module state, shared by both nn.Module and ScriptModule.

abstract forward(ref, inf) → torch.Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

property name

abstract score(pred) → torch.Tensor

espnet2.asvspoof.loss.binary_loss

class espnet2.asvspoof.loss.binary_loss.ASVSpoofBinaryLoss(weight: float = 1.0)

Bases: espnet2.asvspoof.loss.abs_loss.AbsASVSpoofLoss

Binary loss for ASV Spoofing.

forward(pred: torch.Tensor, label: torch.Tensor, **kwargs)

Forward.

Parameters:

• pred (torch.Tensor) – prediction probability [Batch, 2]

• label (torch.Tensor) – ground truth label [Batch, 2]

score(pred: torch.Tensor)

espnet2.asvspoof.loss.__init__

espnet2.asvspoof.loss.am_softmax_loss

class espnet2.asvspoof.loss.am_softmax_loss.ASVSpoofAMSoftmaxLoss(weight: float = 1.0, enc_dim: int = 128, s: float = 20, m: float = 0.5)

Bases: espnet2.asvspoof.loss.abs_loss.AbsASVSpoofLoss

Binary loss for ASV Spoofing.

forward(label: torch.Tensor, emb: torch.Tensor, **kwargs)

Forward.

Parameters:

• label (torch.Tensor) – ground truth label [Batch, 1]

• emb (torch.Tensor) – encoder embedding output [Batch, T, enc_dim]

score(emb: torch.Tensor)

Prediction.

Parameters:

emb (torch.Tensor) – encoder embedding output [Batch, T, enc_dim]

espnet2.asvspoof.loss.oc_softmax_loss

class espnet2.asvspoof.loss.oc_softmax_loss.ASVSpoofOCSoftmaxLoss(weight: float = 1.0, enc_dim: int = 128, m_real: float = 0.5, m_fake: float = 0.2, alpha: float = 20.0)

Bases: espnet2.asvspoof.loss.abs_loss.AbsASVSpoofLoss

Binary loss for ASV Spoofing.

forward(label: torch.Tensor, emb: torch.Tensor, **kwargs)

Forward.

Parameters:

• label (torch.Tensor) – ground truth label [Batch, 1]

• emb (torch.Tensor) – encoder embedding output [Batch, T, enc_dim]

score(emb: torch.Tensor)

Prediction.

Parameters:

emb (torch.Tensor) – encoder embedding output [Batch, T, enc_dim]