新浪科技讯 11月3日下午消息，近日，360集团低调开源的视觉语言对齐模型FG-CLIP2在全球科技圈掀起热议。在涵盖长短文本图文检索、目标检测等在内的29项权威公开基准测试中，这款模型全面超越了科技巨头Google的SigLIP 2与Meta的MetaCLIP2，标志着中国在AI基础模型领域取得了又一突破性进展。
据介绍，360的FG-CLIP2成功攻克了CLIP模型长期存在的“细粒度识别”痛点，面对包含多个物体的复杂场景，其细节识别置信度仍高达96%。在模型核心上，它实现了三大根本创新：第一，层次化对齐架构，让模型能像人眼一样，同时把握宏观场景与微观细节，实现从“看得见”到“看得清”的跨越。第二，动态注意力机制，使模型可以智能聚焦于图像关键区域，以最小算力代价换取精准的细节捕捉能力。第三，双语协同优化策略，从底层解决了中英文理解不平衡的难题，实现了真正的双语原生支持。
责任编辑：刘万里 SF014
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.