kapoorparul

I am getting quite bad utterances when I try to generate audios for just single words instead of sentences using pre trained model. How can I improve those?