Comment by ben_w

Unfortunately, none that I'm aware of. For whatever reason, I find that speech to text is never as good as the accuracy scores claimed by those making the models.