Ahmadzei's picture
added 3 more tables for large emb model
5fa1a76
If the synthesized speech sounds poor, try using a different speaker embedding.