Skip to content

Latest commit

 

History

History
63 lines (44 loc) · 3.55 KB

README.md

File metadata and controls

63 lines (44 loc) · 3.55 KB

EfficientSAM

EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anything

News

[Jan.12 2024] ONNX version of EfficientSAM including separate encoder and decoder is available on the Hugging Face Space (thanks to @wkentaro Kentaro Wada for implementing onnx export)

[Dec.31 2023] EfficientSAM is integrated into the annotation tool, Labelme (huge thanks to lableme team and @wkentaro Kentaro Wada)

[Dec.11 2023] The EfficientSAM model code with checkpoints is fully available in this repository. The example script shows how to instantiate the model with checkpoint and query points on an image.

[Dec.10 2023] Grounded EfficientSAM demo is available on Grounded-Efficient-Segment-Anything (huge thanks to IDEA-Research team and @rentainhe for supporting grounded-efficient-sam demo under Grounded-Segment-Anything).

[Dec.6 2023] EfficientSAM demo is available on the Hugging Face Space (huge thanks to all the HF team for their support).

[Dec.5 2023] We release the torchscript version of EfficientSAM and share a colab.

Online Demo & Examples

Online demo and examples can be found in the project page.

EfficientSAM Instance Segmentation Examples

Point-prompt point-prompt
Box-prompt box-prompt
Segment everything segment everything
Saliency Saliency

Model

EfficientSAM checkpoints are available under the weights folder of this github repository. Example instantiations and run of the models can be found in EfficientSAM_example.py.

EfficientSAM-S EfficientSAM-Ti
Download Download

You can directly use EfficientSAM with checkpoints,

from efficient_sam.build_efficient_sam import build_efficient_sam_vitt, build_efficient_sam_vits
efficientsam = build_efficient_sam_vitt()

Jupyter Notebook Example

The notebook is shared here

Acknowledgement

If you're using EfficientSAM in your research or applications, please cite using this BibTeX:

@article{xiong2023efficientsam,
  title={EfficientSAM: Leveraged Masked Image Pretraining for Efficient Segment Anything},
  author={Yunyang Xiong, Bala Varadarajan, Lemeng Wu, Xiaoyu Xiang, Fanyi Xiao, Chenchen Zhu, Xiaoliang Dai, Dilin Wang, Fei Sun, Forrest Iandola, Raghuraman Krishnamoorthi, Vikas Chandra},
  journal={arXiv:2312.00863},
  year={2023}
}