Step-1o Audio: The Launch of a Trillion-Parameter End-to-End Speech Model Integrated with Yuyuan App

AIbase基地

Published inAI News · 4 min read · Dec 16, 2024

144

Recently, Jieyue Star announced the launch of a new member of the Step series model matrix - the Step-1o end-to-end voice large model, which is said to be the first end-to-end voice large model in the country with over 100 billion parameters. The release of this new model marks a significant advancement in the field of voice technology, enabling integrated voice understanding and generation through an end-to-end voice solution, thus enhancing the model's intelligence and emotional quotient, and providing a supernatural, high-quality communication experience.

The Step-1o model supports mixed input and output forms such as voice and text, allowing for quick responses and interruptions, providing a convenient interactive experience. It can deeply understand and imitate voice characteristics such as tone, rhythm, dialect, and personalized speaking habits, delivering expressions that are lively and emotionally rich, akin to those of a real person. Additionally, Step-1o is a "smart" voice large model that continuously improves response quality through self-learning and imitation, possessing a wealth of knowledge to provide professional advice and emotionally valuable companionship.

WeChat Screenshot_20241216081425.png

The three main features of the Step-1o voice large model include: emotional intelligence online, capable of understanding the complex semantics and emotional information behind human speech and intonation, with a profound understanding of social ethics and rules; intelligence online, able to provide high-quality answers across various professional fields, possessing critical thinking skills; and strong capabilities in understanding, imitation, and creativity, capable of accurately grasping details in voice expression and creating and narrating bizarre voice stories in real-time.

The application range of Step-1o is extensive, supporting the voice interaction technology needs across various industries such as news broadcasting, chat companionship, audiobooks, online education, smart hardware, and automotive sectors.

Recently, Step-1o will be integrated into the Yuewen App, providing real-time voice call services, allowing users to communicate with Step-1o via phone and enjoy its patient listening and strategic advice services. The launch of this new feature will further expand the application scenarios of Step-1o, providing users with a richer voice interaction experience.

State Grid and Alibaba, Baidu Release the 'Bright Power Large Model' with a Trillion Parameters

Recently, State Grid Corporation announced the launch of China's first trillion-parameter artificial intelligence large model in the power industry - the 'Bright Power Large Model'. They have signed a strategic cooperation framework agreement with Baidu and Alibaba. The officials stated that they will work together with the signing parties to build the Bright Power Large Model and promote the integrated development of energy and power technological innovation and industrial innovation.

AI Expert Tan Xu Joins Dark Side of the Moon, Responsible for Developing End-to-End Speech Models

According to media reports, Tan Xu, former Chief Research Manager at Microsoft Research Asia, joined the startup Dark Side of the Moon in August, primarily responsible for developing end-to-end speech models. One important goal after Tan's joining is to create a speech experience similar to GPT-4o for the company. Dark Side of the Moon has not yet responded, but insiders have confirmed Tan's employment, and specific position details have not been finalized. Public information indicates that Tan Xu's research areas include deep learning, language/speech/music processing, and AI content generation, with publications in AI academic conferences.

Westlake-Omni: Open Source Chinese Emotional End-to-End Speech Model

Westlake Xincheng announces the open-source release of its Westlake-Omni model. As the world's first open-source Chinese emotional end-to-end speech interaction model, Westlake-Omni utilizes discrete representation to unify text and speech modalities, with a particular focus on real-time capabilities, enabling quick responses and providing an almost zero-latency experience.

Launched on August 30! iFlytek's Spark Voice Large Model Updates 'Ultra-Human-Like Interaction'

The iFlytek Spark Voice Large Model has received a brand new upgrade, introducing a new generation of interactive experience called 'Ultra-Human-Like Interaction'. This upgrade has optimized and enhanced multiple aspects, aiming to provide users with a more natural, smooth, and emotionally rich dialogue experience.