Is it possible to use a custom model path?

[szkane]

Users who work with voice functionalities generally already have the Qwen TTS model stored locally. It would be much better if we could specify a custom model path, avoiding redundant downloads and excessive storage usage. After designating the directory, the system should scan the files within it and automatically recognize the models.

[JudgeGroovyman]

Thats actually a really good point. I just downloaded this and dont know where it stores them anyway. but there should be an easy way to point it to the right folders