ocean locust
Hey all,
Is there any way for me to use a custom .pth voice model from weights.gg to create a custom voice for Piper on my home assistant? For example, I have found this Spongebob voice that I would like to use (https://www.weights.gg/models/clsx4fmbv001xdq5yqh9ya70l); however the files provided are in the .index and .pth format which is completely different from the .onnx and .onnx.json which Piper uses.
Additionally, I have found this repository which allows for quick voice model training for Piper (https://github.com/domesticatedviking/TextyMcSpeechy), but there is no mention of involving .pth files whatsoever. Does this mean that I would have to manually supply .wav files and corresponding transcripts in a .csv file in order to train my own custom voice for Piper, and the .pth files are a complete deadend?
Any pointers are greatly appreciated. Thank you in advance for any help 🙏
turbid ether