Description
Cutting out every person in red from a video used to mean clicking each one with shaky tracking. SAM 3 is Meta's third-generation Segment Anything Model with text concept prompts, finding every matching object in images and video.
It also takes point, box and exemplar prompts, tracks through video and includes fine-tuning code.
Concepts:Text to objects.
Video:Tracking.
Prompts:Points, boxes and examples.
Fine-tuning:Adapt it.
It also takes point, box and exemplar prompts, tracks through video and includes fine-tuning code.
Features
Concepts:Text to objects.
Video:Tracking.
Prompts:Points, boxes and examples.
Fine-tuning:Adapt it.
