German ASR: Fine-Tuning Wav2Vec2

  • torchaudio.resample is faster than librosa.resample
  • disable group_by_length if there’s a long delay before training starts
    • Made no difference to the outcome

ASR Systems as Models of Phonetic Category Perception in Adults

PHONEME TRANSPOSITION AND TEMPORAL ENCODING IN HUMAN SPEECH RECOGNITION

19th-Century Cockney and RP