Open shigabeev opened 1 year ago
I would really love to had some directions on how to train my model and learn more about this technology.... i could definitely get into the trouble creating datasets, but i have no clue after that how to start and the technicals, can someone provide some wikis docs or directions on where to start for training ?
Since I don't work for Mycroft anymore and can't be sure how long I will be able to access these repos, I've continued my TTS work here: https://github.com/rhasspy/larynx2
Since I don't work for Mycroft anymore and can't be sure how long I will be able to access these repos, I've continued my TTS work here: https://github.com/rhasspy/larynx2
So, MyCroft doesn't provide training scripts ? i tried to use mycroft's .onnx with larynx but get "nlohmann::json_abi_v3_11_2::detail::type_error" i thought .onnx are universal type of models.
Larynx is not compatible with Mimic 3 models. Onnx models are not universal in that sense.
Hi! I've been watching your TTS systems for a while and appreciate your work! So thank you for open-sourcing your code, I genuinely believe that gruut has a great future and would become a standard in modern TTS systems.
I want to ask how to train new models for Mimic3? I like how English sounds, but Russian has a faulty alignment, and I want to address that. I can contribute to your code with an improved version of Russian TTS, but I can only find your GlowTTS training code for the larynx, not mimic3. Is it available, or is it only for personal use?