| --- |
| license: apache-2.0 |
| tags: |
| - kornia |
| - image-segmentation |
| --- |
| |
| # kornia/mobile_sam |
| |
| Pretrained weights for **MobileSAM**, |
| used by [`kornia.models.SegmentAnything`](https://kornia.readthedocs.io/en/latest/models.html). |
| |
| MobileSAM replaces the heavy ViT-H image encoder of SAM with a distilled TinyViT-5M |
| encoder, reducing model size by 60× while retaining most of SAM's segmentation quality. |
| arXiv 2023. |
| |
| **Original repo:** [ChaoningZhang/MobileSAM](https://github.com/ChaoningZhang/MobileSAM) |
| |
| ## Weights |
| |
| | File | Description | |
| |------|-------------| |
| | `mobile_sam.pt` | TinyViT-5M encoder + SAM decoder | |
|
|
| ## Citation |
|
|
| ```bibtex |
| @article{zhang2023mobilesam, |
| title = {Faster Segment Anything: Towards Lightweight SAM |
| for Mobile Applications}, |
| author = {Zhang, Chaoning and Han, Dongshen and Qiao, Yu |
| and Kim, Jung Uk and Bae, Sung-Ho and Lee, Seungkyu |
| and Hong, Choong Seon}, |
| journal = {arXiv preprint arXiv:2306.14289}, |
| year = {2023} |
| } |
| ``` |
|
|
