tuanh123789 / train_hifigan_xtts Goto Github PK

Hello. maybe someone have seen or has pure XTTSv2 training codes, without coqui's spaghetti codes, that crappy coqui's dependency Trainer and so on?
This repo is nice, we can clearly interact with all needed staff
It's kinda hard to train and modify a model, when you model is surrounded by so much useless staff like in coqui/TTS case

Hi! First, i want to thank you for making this repo and sharing your experiences.
I am trying to train XTTS model for Persian language. I read the topics in this repo coqui-ai/TTS#3704 and i found that we have 3 steps to get XTTS model: DVAE -> GPT-2 -> HifiGAN
According to the discussions in that repo, I realized that DVAE does not need fine tuning because it is independent of the language type. Assuming this claim is true, the next step is fine-tuning GPT-2 . Does the rope you provide (this repo), includes fine-tuning GPT-2 too?

All trained hifigan models come out sounding like this. It just generates straight mel spectrogram bands.

https://vocaroo.com/18YQzfRyOJMV

Additionally, I want to ask why some losses increase during training. Should I choose the last checkpoint or the best checkpoint?