Video about the paper:
Open-source implementation (MIT):
I havn’t had a chance to try it out yet, but I may get around to it!
1 Like
I’ve tried it. The results were only fair for the voices I tried, namely Donald trump, Elmer fudd, and Terry Gross.
1 Like
Whilst it may not be something ready for production use, it’s a pretty exciting (or scary) development to see what could be possible in the not too distant future.
Currently if a person who is losing their speech due to a medical condition wants to clone their voice so they can continue speaking like themselves, it costs significant sums. I’m excited by these advancements in transfer learning so that everyone can have a voice that represents them.
2 Likes