SpeechRequest

{
    "model": "cosyvoice-v2",
    "voice": "longxiaochun",
    "input": "string",
    "response_format": "mp3",
    "speed": 1,
    "instructions": "string"
}