947af3a525
- Restructured the navigation in mkdocs.yml to improve organization, introducing subcategories for assistant creation and component libraries. - Added new documentation for workflow configuration options, detailing setup and best practices. - Introduced new sections for voice recognition and generation, outlining configuration items and recommendations for optimal performance.
25 lines
620 B
Markdown
25 lines
620 B
Markdown
# 语音识别
|
||
|
||
语音识别（ASR）负责将用户音频实时转写为文本，供对话引擎理解。
|
||
|
||
## 配置项
|
||
|
||
| 配置项 | 说明 |
|
||
|---|---|
|
||
| ASR 引擎 | 选择语音识别服务提供商 |
|
||
| 模型 | 识别模型名称 |
|
||
| 语言 | 中文/英文/多语言 |
|
||
| 热词 | 提升特定词汇识别准确率 |
|
||
| 标点与规范化 | 是否自动补全标点、文本规范化 |
|
||
|
||
## 建议
|
||
|
||
- 客服场景建议开启热词并维护业务词表
|
||
- 多语言场景建议按会话入口显式指定语言
|
||
- 对延迟敏感场景优先选择流式识别模型
|
||
|
||
## 相关文档
|
||
|
||
- [语音配置总览](voices.md)
|
||
|