近日，上海人工智能实验室（上海AI实验室）携手清华大学、香港中文大学、商汤科技等机构，共同发布了新一代书生·视觉大模型（InternVL）。这一突破性的技术成果，标志着视觉核心任务在开源领域的领先地位进一步巩固。
新一代书生·视觉大模型的视觉编码器参数量高达60亿（InternVL-6B），这一参数量的增加，使得模型在处理复杂视觉任务时更加高效和精准。更为重要的是，该模型首次提出了对比-生成融合的渐进式对齐技术，这一技术能够实现视觉大模型与语言大模型在互联网级别数据上的精细对齐。这一技术的应用，不仅提升了模型的性能，也为后续的研究和应用提供了新的可能性。
上海AI实验室的这一成果，不仅展示了中国在人工智能领域的研究实力，也为全球人工智能技术的发展贡献了中国智慧。随着人工智能技术的不断进步，未来这一技术有望在医疗、教育、自动驾驶等多个领域发挥重要作用，推动社会进步和科技发展。
Title: Shanghai AI Lab Launches Next-Gen Visual Model
Keywords: Visual Model, Open Source Leadership, Progressive Alignment
News content:
Recently, the Shanghai Artificial Intelligence Laboratory (Shanghai AI Lab) in collaboration with Tsinghua University, The Chinese University of Hong Kong, and SenseTime, has launched the latest generation of the Shusheng·Visual Model (InternVL). This groundbreaking technological achievement further solidifies the leading position of visual core tasks in the open-source domain.
The new generation of the Shusheng·Visual Model boasts a visual encoder with a parameter count of 6 billion (InternVL-6B), a significant increase that enables the model to process complex visual tasks more efficiently and accurately. More importantly, the model introduces a novel progressive alignment technique that combines contrastive learning and generative modeling, allowing for fine-grained alignment between visual and language models on internet-scale data. This technique not only enhances the model’s performance but also opens up new possibilities for future research and applications.
The achievements of the Shanghai AI Lab demonstrate China’s research strength in the field of artificial intelligence and contribute Chinese wisdom to the global development of AI technology. As AI technology continues to advance, this technology is expected to play a significant role in various fields such as healthcare, education, and autonomous driving, propelling social progress and technological development.
【来源】https://mp.weixin.qq.com/s/bdfAJRqOF9tUk8Vy9KC_XQ
Views: 1