segment-anything-2-hiera-small-image-torch

Fast image segmentation model that runs efficiently on laptops and edge computing devices.

Details

• Model name: segment-anything-2-hiera-small-image-torch

• Model source: https://ai.meta.com/sam2/

• Model author: Nikhila Ravi, et al.

• Model license: Apache 2.0, BSD 3-Clause

• Model size: 175.77 MB

• Exposes embeddings? no

• Tags: segment-anything, torch, zero-shot, transformer, official

Requirements

• Packages: torch, torchvision, sam2

• CPU support

  • yes

• GPU support

  • yes

Example usage

import fiftyone as fo
import fiftyone.zoo as foz

dataset = foz.load_zoo_dataset(
    "coco-2017",
    split="validation",
    dataset_name=fo.get_default_dataset_name(),
    max_samples=50,
    shuffle=True,
)

model = foz.load_zoo_model("segment-anything-2-hiera-small-image-torch")

# Segment inside boxes
dataset.apply_model(
    model,
    label_field="segmentations",
    prompt_field="ground_truth",  # can contain Detections or Keypoints
)

# Full automatic segmentations
dataset.apply_model(model, label_field="auto")

session = fo.launch_app(dataset)