闪电说本地SenseVoice当前似乎在 sensevoice_onnx.rs 后处理中删除了语言、情绪和声音事件token。希望增加“SenseVoice富文本输出”选项,保留并结构化返回 language、emotion、events 和 text。默认关闭,不影响普通输入;开启后允许在最近记录或自定义技能中访问这些字段。个人语音日记场景需要利用 <|HAPPY|>、<|SAD|>、<|ANGRY|>、<|Laughter|>、<|Breath|> 等模型已经生成的标签,而不只是最终纯文本。
Please authenticate to join the conversation.
In Review
Feature Request
12 days ago

bt parten
Get notified by email when there are changes.
In Review
Feature Request
12 days ago

bt parten
Get notified by email when there are changes.