aligner/README.md
2026-04-17 15:16:12 +08:00

32 lines
435 B
Markdown

# 音频文本对齐服务
本服务部署在有GPU的主机上
## api
请求格式
```
curl -X POST https://server:port/align \
-H "Content-Type: application/json" \
-F "text=音频中的文字" \
-F "audio_file=@/path/to/音频文件"
```
输出:
```
[
{
"sentence": "世界啊你好",
"start": 0.123,
"end": 1.45,
"chars":[
{
"char": "世",
"start":0.123,
"end": 0.543
},
...
]
}
...
]
```