How to create a tts synthesizer with a model non available in coqui ai tts --list_models

[fdemolde]

Hello Rowan,

I trained a fastspeech2 model but I cannot build a synthesizer with the model name in argument.

Error: Result::unwrap()on anErr` value: PyErr { type: <class 'KeyError'>, value: KeyError('fastspeech2'), traceback: Some(<traceback object at 0x752fd3883880>) }

But when I change the name of the model folder (in '.local/share/tts') and the name of the model in the code to a name from tts --list_models, it's working.

I think if I could append the trained model name to this list, it would work.

Does you know how to modify this list locally?

Best regards.