Original text by the authors. Images and external links are unavailable on the intranet.
Hibou-L - a Foundational Vision Transformer for digital pathology pretrained on a 1.2B image private dataset using DINOv2 framework.
For more information and examples of usage visit https://github.com/HistAI/hibou and read the paper.
Hibou-B is also available at https://huggingface.co/histai/hibou-b
Basic usage:
from transformers import AutoImageProcessor, AutoModel
processor = AutoImageProcessor.from_pretrained("histai/hibou-L", trust_remote_code=True)
model = AutoModel.from_pretrained("histai/hibou-L", trust_remote_code=True)
We use a customized implementation of the DINOv2 architecture from the transformers library to add support for registers, which requires the trust_remote_code=True flag.