Struct EcapaTdnn 

pub struct EcapaTdnn { /* private fields */ }

ECAPA-TDNN speaker embedding extractor.

Emphasized Channel Attention, Propagation and Aggregation in TDNN. State-of-the-art for speaker verification (2020+).

Note

This is a stub - actual implementation requires model weights.

Implementations

impl EcapaTdnn

pub fn new(config: EmbeddingConfig) -> Self

Create new ECAPA-TDNN extractor

Note

This is a stub - returns NotImplemented on extract().

pub fn default_config() -> Self

Create with default ECAPA configuration

Trait Implementations

impl Debug for EcapaTdnn

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl EmbeddingExtractor for EcapaTdnn

fn extract(&self, audio: &[f32]) -> VoiceResult<SpeakerEmbedding>

Extract speaker embedding from audio samples.

fn embedding_dim(&self) -> usize

Get the embedding dimension

fn sample_rate(&self) -> u32

Get the expected sample rate