Meta SAM 2.1 is now available in Amazon SageMaker JumpStart

AWS and Meta have announced the availability of Meta Segment Anything Model (SAM) 2.1 in Amazon SageMaker JumpStart, providing advanced vision segmentation capabilities for machine learning practitioners.

• Offers state-of-the-art image and video segmentation with multiple model variants (Tiny, Small, Base Plus, Large)
• Supports zero-shot object segmentation and tracking using simple prompts like point coordinates
• Can be deployed through SageMaker Studio or Python SDK with easy-to-use deployment options
• Provides three primary tasks: automatic mask generator, image predictor, and video predictor
• Supports use cases in medical imaging, autonomous systems, and computer vision research

The model can be deployed on various GPU instances and offers flexible inference capabilities for complex segmentation scenarios across different industries.