Struct AudioEncoder

pub struct AudioEncoder {
    pub config: AudioConfig,
    pub audio_token_id: u32,
    pub begin_audio_token_id: u32,
}

Encoder for converting audio data into token sequences.

The AudioEncoder processes audio waveforms and converts them into token sequences that can be mixed with text tokens in multimodal applications.

Fields

• config - Audio processing configuration
• audio_token_id - Token ID (u32) for audio content tokens
• begin_audio_token_id - Token ID (u32) for marking the start of audio

Fields

config: AudioConfig

audio_token_id: u32

begin_audio_token_id: u32

Implementations

impl AudioEncoder

pub fn new( config: AudioConfig, audio_token_id: u32, begin_audio_token_id: u32, ) -> Self

Creates a new AudioEncoder.

Arguments

• config - Audio processing configuration
• audio_token_id - Token ID (u32) representing audio content
• begin_audio_token_id - Token ID (u32) marking the start of audio sequence

Returns

A new AudioEncoder instance.

pub fn encode(&self, audio: Audio) -> Result<AudioEncoding>

Encodes audio data into a token sequence.

This method processes the audio through resampling, padding, and tokenization to produce a sequence of tokens that represents the audio content.

Arguments

• audio - The audio data to encode

Returns

An AudioEncoding containing the token sequence and processed audio.

Errors

Returns an error if audio processing fails.

Examples

use tekken::audio::{Audio, AudioConfig, AudioSpectrogramConfig, AudioEncoder};

let audio = Audio::from_file("audio.wav")?;
let spectrogram_config = AudioSpectrogramConfig::new(80, 160, 400)?;
let audio_config = AudioConfig::new(16000, 12.5, spectrogram_config, None)?;
let encoder = AudioEncoder::new(audio_config, 1000, 1001);

let encoding = encoder.encode(audio)?;
println!("Audio encoded to {} tokens", encoding.tokens.len());

Trait Implementations

impl Clone for AudioEncoder

fn clone(&self) -> AudioEncoder

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for AudioEncoder

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

Formats the value using the given formatter.