dinov2-vitb14-torch#

Creates searchable image fingerprints for finding similar pictures and organizing large photo collections.

Details

  • Model name: dinov2-vitb14-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 330.33 MB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitb14-torch")
13
14embeddings = dataset.compute_embeddings(model)