[BUG]: Don't use this it's purposefully broken

[VythanLe]

The ReadMe.txts have links that 404 and have not been upkept.
The .sh file "meant to download dependencies" does nothing
The example pages in the readme.txts under the specified egs/models/readme.txt all 404.

downloading the dependencies from hugging face freezes on one bulk of 6 files at 3 files no matter how many times you try it and refuses to download the remaining 3.

there are over 100k files in the VENV created for apparently no reason.

there's two folders for apparently 2 different models:
/pretrained (containing the original models)
/models (containing new added models)

problem is, the pretrained models have no actual models, only training data.
the /models folder contains no training data.

If this even works, whatever idiot is behind it has PURPOSEFULLY destroyed all documentation forcing you to learn every script by heart and brute force it, and even then I am sure they have purposefully broken it as the TTS/Base/TTS model folder DOES NOT EVEN CONTAIN A MODEL.MAIN, only the modules used for model.main.

This has turned into some kids scam, instead you should write your own. All it does is use VAE and GAN according to documentation. If you are going to download it just download it with ripping off the models in mind. Aparently all it does is generate a HUMAN AUDIBLE "mel" audio file using trained data that comes from VAE with added GAN because the GAN apparently "smooths" the audio. All models are just slight parameter variations of this basic concept. You can just code one model and give it the ability to read from different training data that has different parameters on it. I highly suggest you do that instead because this 100k file VENV spaghetti code mess has had it's documentation purposefully removed so that whoever is running it can bank off it (probably running his own TTS website).