中国电信在1月10日宣布开源星辰语义大模型TeleChat-7B版本，并开放1T清洗数据集，以推动开源大模型生态的建设。此外，中国电信还计划在1月20日开源12B版本模型，进一步扩大开发者群体。星辰语义大模型由中国电信人工智能科技有限公司研发训练，采用了1.5万亿Tokens的中英文语料进行训练，旨在提供更精准的语义理解和语言生成能力。这一举措将有助于加速人工智能技术的创新和应用，为开发者提供了强大的工具和支持。
Title: China Telecom Open Sources Starry-Semantics Model
Keywords: China Telecom, Open Source Model, Semantics Model
News content:
China Telecom has announced the open sourcing of the Starry-Semantics Model TeleChat-7B version, along with the release of a 1T cleaned dataset. This move aims to foster the development of an open source large model ecosystem. The company plans to open source the 12B version model on January 20th, inviting more developers to join. The Starry-Semantics Model, developed and trained by China Telecom Artificial Intelligence Technology Co., Ltd., utilizes 1.5 trillion tokens of Chinese and English text data to provide more accurate semantic understanding and language generation capabilities. This initiative is expected to accelerate the innovation and application of AI technology, offering developers powerful tools and support.
【来源】https://www.ithome.com/0/744/969.htm
Views: 2