AI Tools.

Search

image feature extraction by facebook

dinov2-large

DINOv2 ViT-L is Meta's large-scale self-supervised vision transformer, offering significantly better visual representations than the base variant at 4x the parameter count. It achieves near-supervised performance on linear probing for ImageNet.

Summary text generated by an automated pipeline from the model card · Not individually reviewed or run by us · How this page is made

From the model card

Fields below are copied from the tags and counters on the HuggingFace repository facebook/dinov2-large at our last fetch. They are set by the uploader, not verified by us; rows with no tag are omitted. How this page is made.

Publisher (HF namespace)
facebook
Pipeline tag
image-feature-extraction
Library
Transformers
Framework tags
PyTorch
Weight formats
safetensors
License tag
apache-2.0 — read the license file in the repo before relying on it
Papers cited
arXiv:2304.07193
Downloads (HF counter at last fetch)
768,331
Likes (HF counter at last fetch)
117
Model card
https://huggingface.co/facebook/dinov2-large

Use cases

  • High-quality feature extraction for dense prediction tasks (segmentation, depth)
  • Image retrieval in large-scale datasets
  • Backbone for fine-tuning on specialized visual domains
  • Benchmarking self-supervised representation quality

Pros

  • Substantially outperforms ViT-B on linear probing and transfer tasks
  • Self-supervised training — no label dependency during pretraining
  • Apache-2.0 licensed
  • Available as part of a consistent ViT scale ladder (S/B/L/G)

Cons

  • ViT-L requires ~1.3GB weights and ~5GB VRAM for inference
  • Slower than ViT-B for latency-sensitive applications
  • ViT-G/14 provides further gains if compute allows
  • No text alignment — unsuitable for cross-modal tasks

Tags

transformerspytorchsafetensorsdinov2image-feature-extractiondinovisionarxiv:2304.07193license:apache-2.0endpoints_compatibleregion:us