Recently, DingTalk and the speech team of Tongyi Lab jointly announced the launch of a new generation of speech recognition large model - Fun-ASR. This model has excellent speech recognition capabilities, able to understand professional terms in ten industries such as home decoration and animal husbandry, and supports custom training of dedicated models according to enterprise needs.
Fun-ASR was developed based on the deep collaboration between the two parties, capable of accurately transcribing various voice signals, with the ability to recognize multi-industry terminology, handle different languages and accents. In addition, Fun-ASR can perform more accurate speech transcription by combining contextual information, fully meeting the needs of enterprises in specific scenarios. To improve the flexibility and targeting of use, Fun-ASR allows enterprises to customize training of the model according to their own business characteristics, thus better adapting to their respective business environments.
Currently, Fun-ASR has been successfully integrated into multiple functional modules of DingTalk, including meeting subtitles and simultaneous interpretation, smart minutes, and voice assistants. The launch of these features aims to provide enterprises with a stable, efficient, and scalable speech recognition foundation, especially suitable for enterprise scenarios that require high context understanding and recognition accuracy.
Although Fun-ASR currently has considerable functionality, its potential has not been fully explored. In the future, DingTalk and Tongyi Lab will continue to explore and upgrade in areas such as dialect recognition, noise robustness, multilingual support, and enterprise deep customization. Through these continuous efforts, both parties hope to further enhance the accuracy and practicality of speech transcription, promoting the intelligent upgrade of more enterprise business scenarios.
Key Points:
- 🎤 The Fun-ASR speech recognition large model launched by DingTalk and Tongyi Lab is capable of understanding multi-industry terminology.
- 📊 Fun-ASR has been integrated into various functions of DingTalk, supporting enterprise customized training.
- 🚀 Both parties will continue to enhance the capabilities of Fun-ASR in areas such as dialect recognition and noise robustness in the future.
