测试了一下,效果不错,可以用赠金
接口同openAI的tts
模型名和声音需要使用下面的这些
model: “RVC-Boss/GPT-SoVITS”
response_format: “mp3”
voice: “fishaudio/fish-speech-1.4:claire”
fishaudio/fish-speech-1.4:alex, fishaudio/fish-speech-1.4:anna,
fishaudio/fish-speech-1.4:bella, fishaudio/fish-speech-1.4:benjamin,
fishaudio/fish-speech-1.4:charles, fishaudio/fish-speech-1.4:claire,
fishaudio/fish-speech-1.4:david, fishaudio/fish-speech-1.4:diana
最后,貌似这个模型也支持自己上传参考声音。。。参考声音这部分我还没测,但是理论上应该也可以