上海的陆家嘴

中国知名科技公司昆仑万维近日发布了一项名为“天工 2.0”的混合专家模型(MoE)大语言模型。这一模型支持超过15万个汉字上下文窗口,显著提升了语言理解和生成能力。值得注意的是,这是国内首次面向普通消费者(C端用户)免费开放千亿级参数的大语言模型应用。

昆仑万维的这一举措标志着中国在大规模语言模型技术方面迈出了重要一步,与国内外同行相比,显示出中国在人工智能领域的快速发展和雄心。用户可从即日起,在各大手机应用市场下载“天工 AI 智能助手”App,并开始体验“天工 2.0”MoE 大模型的强大功能。

“天工 2.0”的推出,不仅代表了昆仑万维在大模型技术上的突破,也体现了该公司对开放共享科研成果的承诺。该模型预计将为用户带来更为精准、自然的语言交互体验,同时也将助力于各类文本生成、翻译、问答等AI应用的发展。

英文翻译:
Kunlunwanwei Unveils “TianGong 2.0” MoE Large Model, Boosting AI Intelligent Assistant
Keywords: Kunlunwanwei, TianGong 2.0, MoE Large Model

News content:
Chinese technology company Kunlunwanwei has recently launched a mixed expert model (MoE) large language model named “TianGong 2.0.” This model supports more than 150,000 Chinese characters in context windows, significantly enhancing language understanding and generation capabilities. It’s worth noting that this is the first time in China that a billion-level parameter large language model application has been opened to the general consumer (C-end users) free of charge.

This move by Kunlunwanwei marks an important step in China’s大规模语言模型技术, showing the country’s rapid development and ambitions in the field of artificial intelligence, compared with domestic and international peers. Users can download the “TianGong AI Intelligent Assistant” App from various mobile application markets starting today and begin experiencing the powerful features of the “TianGong 2.0” MoE large model.

The introduction of “TianGong 2.0” not only represents a breakthrough in large model technology by Kunlunwanwei but also demonstrates the company’s commitment to open sharing of scientific research results. The model is expected to bring users more precise and natural language interaction experiences and will also contribute to the development of various AI applications such as text generation, translation, question and answer, etc.

【来源】https://www.ithome.com/0/749/207.htm

Views: 1

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注