search

jinhan / tacotron2-vae

Implementation of "Learning Latent Representations for Style Control and Transfer in End-to-end Speech Synthesis"
BSD 3-Clause "New" or "Revised" License
166 stars 33 forks source link

Speaker ID in inference notebook #3

Open richardburleigh opened 5 years ago

richardburleigh commented 5 years ago

Thank you for this amazing implementation. Everything is working perfectly except I wasn't able to figure out how to select speakers during inference (I have trained with 5 speakers).

Any hints would be really appreciated!

indexalice commented 4 years ago

Hello! Sorry for bothering you. I am also trying to implement multi-speakers TTS using IEMOCAP database. Did you train it successfully without changing anything like hparams.py? My net training just cannot work at the beginning. ; ; Any replies would be really appreciated!