metadata
license: cc-by-nc-nd-4.0
language:
- en
Model Card for StainNet
StainNet is a lightweight foundation model for special staining histology images.
The model is a Vision Transformer Small/16 with DINO [1] self-supervised pre-training on 1,418,938 patch images from 20,231 special staining whole slide images (WSIs) in HISTAI [2].
Using StainNet to extract features from special staining pathology image
import timm
import torch
model = timm.create_model('hf_hub:JWonderLand/StainNet', pretrained=True)
preprocess = transforms.Compose([
transforms.Resize(224, interpolation=transforms.InterpolationMode.BICUBIC),
transforms.ToTensor(),
transforms.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5)),
])
model = model.to('cuda')
model.eval()
input = torch.randn([1, 3, 224, 224]).cuda()
with torch.no_grad():
output = model(input) # [1, 384]
Citation
If StainNet is helpful to you, please cite our work.
@misc{TBA
}
References
[1] Caron, M., Touvron, H., Misra, I., Jégou, H., Mairal, J., Bojanowski, P., & Joulin, A. (2021). Emerging properties in self-supervised vision transformers. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 9650-9660).
[2] Nechaev, D., Pchelnikov, A., & Ivanova, E. (2025). HISTAI: An Open-Source, Large-Scale Whole Slide Image Dataset for Computational Pathology. arXiv preprint arXiv:2505.12120.