科大讯飞发布星火语音大模型，超越OpenAI Whisper
今天下午，科大讯飞在发布会上宣布推出星火语音大模型。该模型在多语种语音合成方面取得突破，在首批37个主流语种上整体超越了OpenAI公司推出的Whisper-large-v3，保持了科大讯飞智能语音技术的国际领先水平。
据了解，星火语音大模型在首批40个语种平均MOS分绝对提升0.25，达到4.5，拟人度超过83%。MOS（Mean Opinion Score）是衡量语音合成质量的重要指标，数值越高表示语音合成效果越好。
科大讯飞表示，星火语音大模型的推出将进一步提升其智能语音产品和服务的性能，为用户提供更加自然流畅的语音交互体验。目前，星火语音大模型已应用于科大讯飞的智能客服、智能家居、智能车载等多个领域。
Whisper是OpenAI公司开发的多模态人工智能模型，以其强大的语音识别和翻译能力而闻名。Whisper-large-v3是Whisper模型的最新版本，在语音识别和翻译任务上取得了显著的进步。
科大讯飞在语音识别和合成领域拥有深厚的技术积累，其语音识别技术已广泛应用于智能手机、智能家居、智能车载等多个领域。此次发布的星火语音大模型进一步巩固了科大讯飞在语音技术领域的领先地位。
业内人士表示，星火语音大模型的推出将加速语音合成技术的发展，为人工智能语音交互领域带来新的突破。
英语如下：
**Headline: iFLYTEK’s Xunfei Starfire Large Model SurpassesOpenAI Whisper**
**Keywords:** Speech large model, Surpassing OpenAI, High human likeness
**News Content:** iFLYTEK Releases Xunfei Starfire Speech Large Model, Surpassing OpenAI Whisper
This afternoon, iFLYTEK announced the launch of the Xunfei Starfire speech large model at a press conference. The model has made breakthroughs in multilingual speech synthesis, surpassing OpenAI’s Whisper-large-v3 on the first batch of 37 mainstreamlanguages, maintaining iFLYTEK’s international leadership in intelligent speech technology.
It is understood that the Xunfei Starfire speech large model has an absolute increase of 0.25 in the average MOS (Mean Opinion Score) score on the first batch of 40 languages, reaching 4.5, with a human likeness of over 83%. MOS is an important indicator for measuring the quality of speech synthesis, with higher values indicating better speech synthesis effects.
iFLYTEK said that the launch of the Xunfei Starfire speech large model will further enhance the performance of its intelligent speech products and services, providing userswith a more natural and smooth speech interaction experience. At present, the Xunfei Starfire speech large model has been applied to multiple fields such as iFLYTEK’s intelligent customer service, smart home, and intelligent vehicle.
Whisper is a multimodal AI model developed by OpenAI, known for its powerful speech recognition and translation capabilities. Whisper-large-v3 is the latest version of the Whisper model, which has made significant progress in speech recognition and translation tasks.
iFLYTEK has deep technical accumulation in the field of speech recognition and synthesis, and its speech recognition technology has been widely used in multiple fields such as smartphones, smart homes, and intelligent vehicles. The Xunfei Starfire speech large model released this time further consolidates iFLYTEK’s leading position in the field of speech technology.
Industry insiders said that the launch of the Xunfei Starfire speech large model will accelerate the development of speech synthesis technology and bring new breakthroughs to the field of AI speech interaction.
【来源】https://news.mydrivers.com/1/961/961266.htm
Views: 3