Overall recognition quality
未关闭
- 主要语言
- Jupyter Notebook
- 星标
- 15.1k
- 派生
- 1.8k
- PR 合并指标
- 30 天内没有已合并 PR
描述
I used Kaldi recognizer on my voice assistant (Mycroft.ai) for two years. When I setted all things up I checked deepspeech and kaldi with some models found around internet, and found Kaldi better fitted my needs. I tuned word detection probabilities and it somehow works - at least for "switch on light" and "set timer for three minutes". And this month I decided to check vosk, set it up, and ... it works. Without tuning, I just talking and it recognizes it all, almost instantly. It's pure magic! Just wanted to thank you guys for your amazing work! Keep going!
贡献指南
这个仓库没有索引到贡献指南
评估
这个 Issue 还没有评估数据。