近日，阿里云宣布其多模态大模型研究取得重大进展。据悉，通义千问视觉理解模型Qwen-VL再次升级，推出了Max版本。这一升级版模型不仅拥有更强的视觉推理能力和中文理解能力，还能根据图片识人、答题、创作、写代码等。在多个权威测评中，Qwen-VL-Max表现出色，整体性能可与GPT-4V和Gemini Ultra相媲美。
在MMMU、MathVista等测评中，Qwen-VL-Plus和Qwen-VL-Max的成绩远超业界所有开源模型。尤其在文档分析（DocVQA）、中文图像相关（MM-Bench-CN）等任务上，Qwen-VL-Max的表现更是超越GPT-4V，达到了世界最佳水平。
阿里云的这一突破性成果，无疑为我国人工智能领域的发展增添了浓墨重彩的一笔。作为我国领先的云计算企业，阿里云在人工智能领域的持续投入和创新，将进一步推动我国人工智能技术的发展，助力我国科技事业迈向更高的峰。
Title: Alibaba Cloud Qwen-VL Visual Understanding Model Achieves New Milestone
Keywords: Alibaba Cloud, Qwen-VL, Visual Understanding Model
News content:
Recently, Alibaba Cloud announced significant progress in its multi-modal large model research. It is learned that the Tsinghua University KEG Lab Visual Understanding Model Qwen-VL has been upgraded again, with the launch of the Max version. This upgraded version not only has stronger visual reasoning and Chinese understanding capabilities but is also capable of identifying people, answering questions, creating, and coding based on images. In multiple authoritative evaluations, Qwen-VL-Max has shown remarkable performance, with overall capabilities comparable to GPT-4V and Gemini Ultra.
In evaluations such as MMMU and MathVista, Qwen-VL-Plus and Qwen-VL-Max achieved scores far beyond all industry open-source models.特别是在文档分析（DocVQA）、中文图像相关（MM-Bench-CN）等任务上，Qwen-VL-Max’s performance surpassed GPT-4V, reaching the world’s best level.
This breakthrough achievement by Alibaba Cloud has undoubtedly added a brilliant touch to the development of China’s artificial intelligence field. As China’s leading cloud computing company, Alibaba Cloud’s continuous investment and innovation in artificial intelligence will further promote the development of China’s artificial intelligence technology and help China’s science and technology community reach even greater heights.
【来源】https://news.mydrivers.com/1/960/960575.htm
Views: 1