# segment-anything-3-video-torch Open-vocabulary video segmentation that finds, segments, and tracks all objects matching a text concept across video frames. **Details** - Model name: `segment-anything-3-video-torch` - Model source: [https://github.com/facebookresearch/sam3](https://github.com/facebookresearch/sam3) - Model author: Nicolas Carion, Laura Gustafson, Yuan-Ting Hu, et al. - Model license: SAM License - Model size: 3.45 GB - Exposes embeddings? no - Tags: `segment-anything, torch, zero-shot, video, transformer, official` **Requirements** - Packages: `torch, torchvision, sam3` - CPU support - yes - GPU support - yes **Example usage** ```python import fiftyone as fo import fiftyone.zoo as foz from fiftyone import ViewField as F dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2) # Concept mode: find and segment object with text prompts on selected frame, and track objects across frames model = foz.load_zoo_model( "segment-anything-3-video-torch", classes=["person"], operation_mode="concept", propagation_direction="forward", # also supports backward and both text_frame_idx=1, ) dataset.apply_model(model, label_field="segmentations_concept") # Exemplar concept mode: text prompt + exemplar boxes on frame 10, propagate in both directions model = foz.load_zoo_model( "segment-anything-3-video-torch", classes=["person"], operation_mode="concept", propagation_direction="both", prompt_frame_indices=[10], ) dataset.apply_model( model, label_field="segmentations_concept_with_exemplar", prompt_field="frames.person_detections", # exemplar Detections on frame 10 ) # Visual mode: segment inside boxes and propagate to all frames model = foz.load_zoo_model("segment-anything-3-video-torch") dataset.apply_model( model, label_field="segmentations", prompt_field="frames.detections", # can contain Detections or Keypoints prompt_frame_indices=[1], ) session = fo.launch_app(dataset) ```