中国人工智能技术公司零一万物近日宣布其Yi Vision Language（Yi-VL）多模态语言大模型已正式面向全球开源。该模型基于Yi语言模型开发，旨在通过结合图像和文本处理能力，进一步推动自然语言处理技术的发展。Yi-VL模型包括两个版本：Yi-VL-34B和Yi-VL-6B，它们在英文数据集MMMU和中文数据集CMMMU上均取得了领先成绩，展现了其在复杂跨学科任务上的强大实力。这一开源模型的发布，不仅为全球开发者提供了强大的工具，也是中国AI技术在多模态研究领域的一个重大进展。
Title: Yi-VL Multimodal Language Model Open-Sourced Globally
Keywords: Language Model, Open-Source, Cross-Disciplinary Tasks
News content:
ZeroOne Everything, an AI technology company in China, has recently announced the global open-source release of its Yi Vision Language (Yi-VL) multimodal language large model. Developed based on the Yi language model, Yi-VL is designed to advance natural language processing technology by combining image and text processing capabilities. The Yi-VL model includes two versions: Yi-VL-34B and Yi-VL-6B, which have both achieved leading results on the English dataset MMMU and the Chinese dataset CMMMU, demonstrating its powerful capabilities in complex cross-disciplinary tasks. The release of this open-source model provides a powerful tool for developers worldwide and marks a significant advancement of China’s AI technology in the field of multimodal research.
【来源】https://www.jiqizhixin.com/articles/2024-01-22-10
Views: 1