Model← All models
Segment Anything (SAM): promptable image segmentation
Give SAM a point or a box on an image and it returns a mask for that object, or let it mask everything automatically. It handles new image domains well without retraining.
| Organization | Meta AI (FAIR) |
|---|---|
| Released | 2023 |
| Sizes | ViT-B, ViT-L, ViT-H backbones |
| License | Apache-2.0 (code and weights) |
- A heavy image encoder runs once per image, then a light decoder answers each prompt in milliseconds
- Mask decoder can be exported to ONNX for use in the browser
- Note: the SA-1B training dataset has its own separate research license
Summary written by SolxLab. Weights and code belong to their authors and are distributed under the license shown; always check the official page before use.
FrameworkPyTorch
Vision