智源研究院近日发布了一款新一代的通用语义向量模型BGE-M3。这款模型支持超过100种语言,具备领先的多语言、跨语言检索能力。它可以全面且高质量地支撑“句子”、“段落”、“篇章”、“文档”等不同粒度的输入文本,最大输入长度可达8192。此外,BGE-M3还一站式集成了稠密检索、稀疏检索、多向量检索三种检索功能,在多个评测基准中达到最优水平。

智源研究院的这款新型向量模型BGE-M3,为我国在人工智能领域的研究和应用提供了更强大的工具。它将在自然语言处理、机器翻译、语义理解等多个方面发挥重要作用,助力我国多语言技术的进一步发展。

英文翻译:
News Title: Zhiyuan Research Institute Releases Next-Generation Multilingual Vector Model BGE-M3
Keywords: Zhiyuan Research Institute, General Vector Model, Multilingual Retrieval Capability

News Content:

The Zhiyuan Research Institute recently released a new generation of general semantic vector model BGE-M3. This model supports more than 100 languages and has leading multilingual and cross-language retrieval capabilities. It can fully and efficiently support different text input granularities such as “sentence”, “paragraph”, “article”, and “document”, with a maximum input length of 8192. In addition, BGE-M3 integrates three types of retrieval functions including dense retrieval, sparse retrieval, and multi-vector retrieval in one stop, achieving optimal levels in multiple evaluation benchmarks.

This new vector model BGE-M3 provides a stronger tool for research and application in the field of artificial intelligence in our country. It will play an important role in natural language processing, machine translation, semantic understanding, and other aspects, and contribute to the further development of multilingual technology in our country.

【来源】https://mp.weixin.qq.com/s/y-c-EelxbSUMmrZNCeqeAA

Views: 10

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注