Skip to content

Commit

Permalink
Update websocket_protocol_zh.md
Browse files Browse the repository at this point in the history
  • Loading branch information
lyblsgo authored Jan 17, 2024
1 parent e86630a commit dcc310e
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion runtime/docs/websocket_protocol_zh.md
Original file line number Diff line number Diff line change
Expand Up @@ -67,7 +67,7 @@ message为(需要用json序列化):
```text
`mode`:`offline`,表示推理模式为一句话识别;`online`,表示推理模式为实时语音识别;`2pass`:表示为实时语音识别,并且说话句尾采用离线模型进行纠错。
`wav_name`:表示需要推理音频文件名
`wav_format`:表示音视频文件后缀名,可选pcm、mp3、mp4等(备注:只支持pcm音频流
`wav_format`:表示音视频文件后缀名,只支持pcm音频流
`is_speaking`:表示断句尾点,例如,vad切割点,或者一条wav结束
`chunk_size`:表示流式模型latency配置,`[5,10,5]`,表示当前音频为600ms,并且回看300ms,又看300ms。
`audio_fs`:当输入音频为pcm数据是,需要加上音频采样率参数
Expand Down

0 comments on commit dcc310e

Please sign in to comment.