-
In [Towards End-to-End Prosody Transfer for Expressive Speech Synthesis with Tacotron](https://arxiv.org/pdf/1803.09047.pdf), the authors have implemented multi-speaker models by conditioning on speak…
-
deep mind's new paper "Towards End-to-End Prosody Transfer for Expressive Speech Synthesis with Tacotron" , use new attention "GMM attention" ,they find improves generalization to long utterances, Do …
-
In #126 it is mentioned that most of the ability to clone voices lies in the encoder. @mbdash is contributing a GPU to help train a better encoder model.
* Increase the number of hidden layers to 7…
ghost updated
3 years ago
-
Global Style Tokens are embeddings that capture prosodic styles across the training set. This allows the system to explicitly specify the desired prosody of a generated sequence, i.e. essentially how …
-
Thanks for publishing the code and basic training instructions!
## Environment
**Datasets:** (9,063 speakers)
- LibriTTS (train-other-500)
- VoxCeleb1
- VoxCeleb2
- OpenSLR (42-44, 61-66, 69…
-
link: https://arxiv.org/abs/1711.00520
referenced from:
- Style Tokens: Unsupervised Style Modeling, Control and Transfer in End-to-End Speech Synthesis #63
-
link:
- https://arxiv.org/abs/1803.09047
- https://openreview.net/forum?id=ByZiJi-uWH
referenced from:
- Style Tokens: Unsupervised Style Modeling, Control and Transfer in End-to-End Speech Syn…
-
Hi @erogol, I've been a bit off the radar for the past month because of vacation and other projects, but now I am back and ready for action! I am looking into how to do multi speaker embeddings, and h…
-
Hello!
I Spent some time to realize that there are 2 active in parallel web/digital publication manifests maintained by the same group
Web Publications
https://github.com/w3c/wpub/
https://w3c…
-
When two lines have the same "title" AND "year" AND "authors" keep only one (the first? the most complete?)
Below the three key in ISI and SCOPUS:
- title --> ISI = TI / Scopus = Title
- year -->…
tommv updated
3 years ago