上海人工智能实验室（上海AI实验室）近日联合清华大学、香港中文大学、商汤科技等机构，发布新一代书生·视觉大模型（InternVL），在视觉核心任务上实现开源领先。这款新一代“书生·视觉基础”模型的视觉编码器参数量达到了60亿（InternVL-6B），这也是首次提出了对比-生成融合的渐进式对齐技术，实现了在互联网级别数据上视觉大模型与语言大模型的精细对齐。
此次发布的 InternVL，不仅在技术上取得了重大突破，而且其开源的特性，也使得全球的研究者和开发者都可以自由使用和修改，进一步推动人工智能技术的发展。这也是上海AI实验室一直致力于推动的，即通过开源共享，加速人工智能技术的创新和应用。
这款视觉大模型的发布，也标志着我国在人工智能领域取得了新的重要进展。我国一直致力于人工智能技术的研究和发展，希望通过此次发布的视觉大模型，能够进一步推动我国人工智能技术的发展，实现人工智能技术的广泛应用。
英文标题Title：Shanghai AI Lab Unveils Open-Source Visual Large Model
英文关键词Keywords：AI Lab, Visual Large Model, Open-Source Leadership
英文新闻内容News content：
The Shanghai Artificial Intelligence Laboratory (Shanghai AI Lab) recently joined forces with Tsinghua University, the Chinese University of Hong Kong, and SenseTime to launch a new generation of Scholar Visual Large Model (InternVL), leading the open-source trend in visual core tasks. This new generation of “Scholar Visual Base” model features a visual encoder with 6 billion parameters (InternVL-6B), marking the first time a comparative-generative fusion progressive alignment technology has been proposed, achieving fine alignment of visual large models with language large models on internet-scale data.
This release of InternVL not only represents a significant breakthrough in technology but also marks the Lab’s ongoing commitment to open-source sharing, enabling researchers and developers worldwide to access and modify the model freely, further propelling the advancement of artificial intelligence technology. It aligns with Shanghai AI Lab’s mission to promote innovation and application of AI technology through open-source collaboration.
The launch of this visual large model signifies another key milestone in China’s AI field. China has been dedicated to the research and development of AI technology, aiming to further promote its progress and widespread application through the release of this visual large model.
【来源】https://mp.weixin.qq.com/s/bdfAJRqOF9tUk8Vy9KC_XQ
Views: 1