Segment images with SAM 48.9x faster by using EfficientViT with EfficientViT-SAM
Segment images with SAM 48.9x faster by using EfficientViT with EfficientViT-SAM
EfficientViT-SAM: Accelerated Segment Anything Model Without Accuracy Loss
arXiv paper abstract https://arxiv.org/abs/2402.05008
arXiv PDF paper https://arxiv.org/pdf/2402.05008
GitHub https://github.com/mit-han-lab/efficientvit
… present EfficientViT-SAM, a new family of accelerated segment anything models.
… retain SAM’s lightweight prompt encoder and mask decoder while replacing the heavy image encoder with EfficientViT.
For the training, … begin with the knowledge distillation from the SAM-ViT-H image encoder to EfficientViT.
Subsequently, … conduct end-to-end training on the SA-1B dataset.
Benefiting from EfficientViT’s efficiency and capacity, EfficientViT-SAM delivers 48.9x measured TensorRT speedup on A100 GPU over SAM-ViT-H without sacrificing performance.
… code and pre-trained models are released …
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
