Michael Hansen
Michael Hansen
Yes, the "low quality" lessac checkpoint was accidentally trained with the wrong parameters. I'm retraining it now :slightly_smiling_face: I'd suggest using the lessac medium checkpoint (22050 Hz sample rate).
How many speakers are in your new dataset (specifically `num_speakers` in `config.json`)?
It is possible to train multi-speaker models, but I haven't done any tests with multiple languages in a single model. There's nothing preventing this, though.
Do you know of any Cantonese voice dataset?
I need a TTS (text to speech) dataset which has one person reading a script with a good microphone in a quiet environment. The dataset must also have an open...
What version?
Is this for a 32-bit OS?
Do you think a new module would be able to work in C++ as well as Python?
Development has moved: https://github.com/OHF-Voice/piper1-gpl What do you think about just enabling all of the GPU providers if `--cuda` is passed?
In the next version of Piper, I have actually ported epitran to become part of Piper's phonemizer 🙂