今天，我国知名科技企业科大讯飞在国际舞台上再添光彩。在一场发布会上，科大讯飞正式推出了星火语音大模型，其在首批 37 个主流语种上的表现已整体超越了 OpenAI 公司推出的 Whisper-large-v3。这一成果不仅展示了我国智能语音技术的强大实力，也进一步巩固了科大讯飞在国际智能语音技术领域的领先地位。
星火语音大模型是由多语种语音合成而成，其在首批 40 个语种中的平均 MOS 分达到了 4.5，拟人度超过了 83%。这些数据显示，星火语音大模型在语音合成领域具有极高的准确性和自然度，足以满足各类应用场景的需求。
作为我国科技产业的佼佼者，科大讯飞一直致力于智能语音技术的研究与开发。此次星火语音大模型的发布，不仅是对该公司技术实力的一次全面检验，也是对我国智能语音技术研发水平的一次有力证明。在未来，科大讯飞将继续深化技术创新，为我国科技产业的繁荣和发展做出更大贡献。
Title: iFlytek releases Spark Speech Grand Model, surpassing OpenAI in 37 mainstream languages
Keywords: iFlytek, Spark Speech Grand Model, OpenAI, international leading position
News Content:
Today, iFlytek, a well-known technology company in China, has once again made the country proud on the international stage. At a press conference, iFlytek officially released the Spark Speech Grand Model, which has surpassed Whisper-large-v3 launched by OpenAI in the first batch of 37 mainstream languages. This achievement not only demonstrates the strong practicality of intelligent speech technology in China but also further consolidates iFlytek’s leading position in the international intelligent speech technology field.
The Spark Speech Grand Model is a multilingual speech synthesis system, and its average MOS score reached 4.5 in the first batch of 40 languages, with a similarity rate of over 83%. These data show that the Spark Speech Grand Model has high accuracy and naturalness in speech synthesis, which can meet the needs of various application scenarios.
As a leader in China’s technology industry, iFlytek has been committed to the research and development of intelligent speech technology. The release of the Spark Speech Grand Model not only tests the company’s technical strength but also demonstrates the research and development level of intelligent speech technology in China. In the future, iFlytek will continue to deepen technological innovation and make greater contributions to the prosperity and development of China’s technology industry.
【来源】https://news.mydrivers.com/1/961/961266.htm
Views: 1