3 ms·There are no "pure speech" models. All ASR uses language models.by kleiba 4y agoThere are no "pure speech" models. All ASR uses language models.