Dinov2 Small is a small-sized vision transformer model trained using the DINOv2 method and hosted on Hugging Face. It belongs to the class of image feature extraction models and is provided under the apache-2.0 license by AI at Meta.
The model is implemented as a Vision Transformer and distributed in PyTorch and Safetensors formats. It can be loaded through the Transformers library using either a high-level pipeline for image-feature-extraction or by directly instantiating an AutoImageProcessor and AutoModel. Example code demonstrates importing these classes from the transformers package and calling from_pretrained with the repository identifier facebook/dinov2-small.
The repository lists the model card topics as image feature extraction, transformers, and dino. It references the paper arxiv:2304.07193. The page supplies instructions for use with libraries, inference providers, notebooks, and local apps, including direct links to Google Colab and Kaggle notebooks.
Dinov2 Small is made available for download and integration at no cost under its open-source license.
Dinov2 Small is an Other AI product. It focuses on providing high-quality image feature extraction and representation learning for computer vision tasks. It is built as an open-source project for computer vision researchers. Dinov2 Small is open source under the Apache-2.0 license. The product ships for the web, the command line, and API.
It is developed by Meta AI, and the product first shipped in 2023. Development happens publicly on GitHub with 13.1k stars and 1 commits in the last 90 days. Key capabilities include image feature extraction, vision transformer, and representation learning.
Latest indexed changes and source events
facebook/dinov2-small verified by the PulseGate indexer
Other apps tracked under the same category.
