云从科技从容大模型在多模态评测中全球领先，击败GPT-4V等巨头
近日，权威综合评测平台OpenCompass公布的多模态评测结果中，云从科技的从容大模型表现出色，平均得分65.5，成功跻身全球前三。这一成绩超越了谷歌的Gemini-1.5-Pro和GPT-4v，仅次于GPT-4o（69.9）和Claude3.5-Sonnet（67.9）。
OpenCompass大模型开放评测体系是上海人工智能实验室推出的完整开源可复现的评测框架，其多模态评测采用了8个具有代表性的数据集，从多种视角客观量化多模态大模型的能力，评估维度覆盖多个方面。
在国内市场中，从容大模型的成绩也显著，超越了InternVL-Chat（61.7）和GLM-4V（60.8），排名首位。此项成就标志着云从科技在多模态人工智能领域的突出进步和强大实力。
从容大模型2.0的多模态能力展示出了它在不同领域中的均衡高水平表现，包括目标检测、文字识别、动作识别等。此次在OpenCompass评测中的卓越表现，充分证明了从容大模型的先进性和可靠性。
云从科技凭借其在多模态领域的突破，再次展现了其强大的人工智能实力，并为行业发展树立了新的标杆。期待云从科技在未来继续带来更多创新和突破，推动人工智能行业的发展。
英语如下：
News Title: “Yuncong Technology’s Congrong Big Model Enters the Global Top 3, Multi-modal Capabilities Impress in OpenCompass Evaluation”
Keywords: Yuncong Technology, Multi-modal Big Model, Global Top 3
News Content:
Yuncong Technology’s Congrong Big Model Leads the Global Multi-modal Evaluation, Overcoming Giants Such as GPT-4V
Recently, the authoritative comprehensive evaluation platform OpenCompass released its multi-modal evaluation results, in which Yuncong Technology’s Congrong Big Model showed excellent performance, with an average score of 65.5, successfully ranking in the global top 3. This score surpassed Google’s Gemini-1.5-Pro and GPT-4v, following closely behind GPT-4o (69.9) and Claude3.5-Sonnet (67.9).
The OpenCompass Big Model Open Evaluation System is a complete open-source and reproducible evaluation framework launched by the Shanghai Artificial Intelligence Laboratory. Its multi-modal evaluation adopted eight representative datasets to objectively quantify the capabilities of multi-modal big models from multiple perspectives, covering multiple evaluation dimensions.
In the domestic market, the Congrong Big Model also achieved remarkable results, surpassing InternVL-Chat (61.7) and GLM-4V (60.8) to rank at the top. This achievement marks Yuncong Technology’s outstanding progress and strong capabilities in the field of multi-modal artificial intelligence.
The multi-modal capabilities of Congrong Big Model 2.0 have demonstrated its balanced and high-level performance in different fields, including object detection, text recognition, action recognition, etc. Its excellent performance in the OpenCompass evaluation fully proves the advanced and reliable nature of the Congrong Big Model.
Yuncong Technology has once again demonstrated its strong artificial intelligence capabilities through its breakthroughs in the multi-modal field, setting a new benchmark for industry development. It is expected that Yuncong Technology will continue to bring more innovation and breakthroughs in the future to promote the development of the artificial intelligence industry.
【来源】https://www.jiqizhixin.com/articles/2024-06-28-16
Views: 7