Tencent Hunyuan Hy ASR 3.0 Preview Released: Speech Recognition That Understands Context
Tencent
Tencent has released the preview of its Hunyuan Hy ASR 3.0 speech recognition model, which is capable of understanding context. The model can handle both Chinese and English, and supports features such as automatic punctuation, speaker diarization, and custom vocabulary. It also offers on-premise deployment options and supports long audio segments.
Tencent released a preview of its Hunyuan Hy ASR 3.0 speech recognition model. The new model is designed to understand context, improving recognition accuracy by considering the surrounding words and phrases. It supports both Chinese and English, and can automatically add punctuation and identify different speakers. It also allows users to customize the vocabulary for specific domains. The model can be deployed on-premise for enhanced data privacy, and it supports audio segments up to one hour in length. Tencent also provides updates to the model every one to three months to ensure accuracy.
- Abbreviations
- ASR = Automatic Speech Recognition — автоматическое распознавание речи
Source: Tencent Hunyuan (GNews) —
original