--- title: sam3.1 canonical_url: "https://www.modelscope.cn/models/facebook/sam3.1" md_url: "https://www.modelscope.cn/models/facebook/sam3.1.md" repository: facebook/sam3.1 last_updated: 2026-03-27 license: other model_type: - sam3_video architectures: - Sam3VideoModel library_name: - pytorch frameworks: - pytorch language: - en downloads: 5403 stars: 57 tags: - sam3.1 --- # sam3.1 > sam3.1 - facebook 在 ModelScope 开源的模型。SAM 3.1 SAM 3 (Segment Anything with Concepts) is a unified foundation model from Meta for promptable segmentation in images and videos. It detects, segments, and tracks objects using text or visual prompts such as… facebook/sam3.1 是 ModelScope 魔搭社区上的机器学习模型，采用 other 许可。 - **Repository**: facebook/sam3.1 - **License**: other - **Tags**: sam3.1 - **Downloads**: 5403 - **Stars**: 57 - **Last updated**: 2026-03-27 Source: https://www.modelscope.cn/models/facebook/sam3.1 --- # SAM 3.1 [SAM 3](https://github.com/facebookresearch/sam3) (Segment Anything with Concepts) is a unified foundation model from Meta for promptable segmentation in images and videos. It detects, segments, and tracks objects using text or visual prompts such as points, boxes, and masks. SAM 3 introduces the ability to exhaustively segment all instances of an open-vocabulary concept specified by a short text phrase, handling over 50x more unique concepts than existing benchmarks. SAM 3.1 builds on this with **Object Multiplex**, a shared-memory approach for joint multi-object tracking that delivers ~7x faster inference at 128 objects on a single H100 GPU without sacrificing accuracy, along with improved VOS performance on 6 out of 7 benchmarks.

This repository hosts only the SAM 3.1 model checkpoints — there is no Hugging Face Transformers integration. For installation, code, usage examples, and full documentation, please visit the [SAM 3 GitHub repository](https://github.com/facebookresearch/sam3).