Skip to main content

detect_language

Function detect_language 

Source
pub fn detect_language(
    encoder_output: &[f32],
    encoder_shape: &[usize],
) -> Result<LanguageDetection, SpeechError>
Expand description

Detect language from audio features (G2)

Uses encoder outputs to classify the spoken language. Supports 99 languages following Whisper’s language detection.

§Arguments

  • encoder_output - Encoder hidden states
  • encoder_shape - Shape as [batch, frames, hidden_dim]

§Returns

Language detection result with confidence scores

§Note

This is a placeholder that returns a default result. Real implementation requires language detection head weights.