开源推荐
morning
GitHub 热门项目:sam3
摘要
GitHub项目:sam3 仓库地址:https://github。
https
com
github
scholar
google
2026-05-26
1 阅读
GitHub Trending
GitHub 项目:sam3
仓库地址:https://github.com/facebookresearch/sam3
星级:10170 | 作者:facebookresearch
项目描述:该存储库提供了使用 Meta Segment Anything Model 3 (SAM 3) 运行推理和微调的代码、下载经过训练的模型检查点的链接以及展示如何使用该模型的示例笔记本。
===================================================
自述文件内容:
# SAM 3:用概念分割任何东西
元超级智能实验室
[尼古拉斯·卡里恩](https://www.nicolascarion.com/)\*,
[劳拉·古斯塔夫森](https://scholar.google.com/itations?user=c8IpF9gAAAAJ&hl=en)\*,
[胡元廷](https://scholar.google.com/itations?user=E8DVVYQAAAAJ&hl=en)\*,
[Shoubhik Debnath](https://scholar.google.com/itations?user=fb6FOfsAAAAJ&hl=en)\*,
[胡荣航](https://ronghanghu.com/)\*,
[Didac Suris](https://www.didacsuris.com/)\*,
[Chaitanya Ryali](https://scholar.google.com/itations?user=4LWx24UAAAAJ&hl=en)\*,
[Kalyan Vasudev Alwala](https://scholar.google.co.in/itations?user=m34oaWEAAAAJ&hl=en)\*,
[Haitham Khedr](https://hkkhedr.com/)\*, Andrew Huang,
[杰雷](https://jayleicn.github.io/),
[马腾宇](https://scholar.google.com/itations?user=VeTSl0wAAAAJ&hl=en),
[郭百山](https://scholar.google.com/itations?user=BC5wDu8AAAAJ&hl=en),
阿尔皮特·卡拉 (Arpit Kalla),[马库斯·马克斯](https://damaggu.github.io/),
[约瑟夫·格里尔](https://scholar.google.com/itations?user=guL96CkAAAAJ&hl=en),
王猛,[孙培泽](https://peizesun.github.io/),
[罗马·拉德尔](https://scholar.google.com/itations?user=Tpt57v0AAAAJ&hl=en),
[Triantafyllos Afouras](https://www.robots.ox.ac.uk/~afourast/),
[Effrosyni Mavroudi](https://scholar.google.com/itations?user=vYRzGGEAAAAJ&hl=en),
[徐凯](https://k8xu.github.io/)°,
[吴宗汉](https://patrickthwu.com/)°,
[于舟](https://yu-bryan-zhou.github.io/)°,
[Liliane Momeni](https://scholar.google.com/itations?user=Lb-KgVYAAAAJ&hl=en)°,
[Rishi Hazra](https://rishihazra.github.io/)°,
[丁双锐](https://mark12ding.github.io/)°,
[Sagar Vaze](https://sgvaze.github.io/)°,
[弗朗索瓦·波尔彻](https://scholar.google.com/itations?user=LgHZ8hUAAAAJ&hl=en)°,
[冯立](https://fengli-ust.github.io/)°,
[李思源](https://siyuanliii.github.io/)°,
[艾西瓦娅·卡马斯](https://ashkamath.github.io/)°,
[郑浩其](https://hk Chengrex.com/)°,
[Piotr Dollar](https://pdollar.github.io/)†,
[尼基拉·拉维](https://nikhilaravi.com/)†,
[凯特·萨恩科](https://ai.bu.edu/ksaenko.html)†,
[张鹏川](https://pzzhang.github.io/pzzhang/)†,
[Christoph Feichtenhofer](https://feichtenhofer.github.io/)†
\* 核心贡献者,° 实习生,† 项目负责人,组内顺序随机
[[`论文`](https://ai.meta.com/research/publications/sam-3-segment-anything-with-concepts/)]
[[`项目`](https://ai.meta.com/sam3)]
[[`演示`](https://segment-anything.com/)]
[[`博客`](https://ai.meta.com/blog/segment-anything-model-3/)]
[[`BibTeX`](#citing-sam-3)]
 SAM 3 是一个统一的基础模型,用于快速分割图像和视频。它可以使用文本或视觉提示(例如点、框和蒙版)来检测、分割和跟踪对象。与其前身 [SAM 2](https://github.com/facebookresearch/sam2) 相比,SAM 3 引入了对由短文本短语或范例指定的开放词汇概念的所有实例进行详尽分段的能力。与之前的工作不同,SAM 3