Model card for vit_base_patch16_224.owkin_pancancer_ft_lc25000_colon

A Vision Transformer (ViT) image classification model.
Trained by Owkin on 40M pan-cancer histology tiles from TCGA.
Fine-tuned on LC25000's colon subset.

Model Details

• Model Type: Image classification / feature backbone
• Model Stats:
  • Params (M): 85.8
  • Image size: 224 x 224 x 3
• Papers:
  • Scaling Self-Supervised Learning for Histopathology with Masked Image Modeling: https://www.medrxiv.org/content/10.1101/2023.07.21.23292757v2
• Pretrain Dataset: TGCA: https://portal.gdc.cancer.gov/
• Dataset: LC25000: https://huggingface.co/datasets/1aurent/LC25000
• Original: https://github.com/owkin/HistoSSLscaling/
• License: https://github.com/owkin/HistoSSLscaling/blob/main/LICENSE.txt

Model Usage

Image Classification

from urllib.request import urlopen
from PIL import Image
import timm

# get example histology image
img = Image.open(
  urlopen(
    "https://datasets-server.huggingface.co/cached-assets/1aurent/LC25000/--/56a7c495692c27afd294a88b7aadaa7b79d8e270/--/default/train/24999/image/image.jpg"
  )
)

# load model from the hub
model = timm.create_model(
  model_name="hf-hub:1aurent/vit_base_patch16_224.owkin_pancancer_ft_lc25000_colon",
  pretrained=True,
).eval()

# get model specific transforms (normalization, resize)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0))  # unsqueeze single image into batch of 1

Image Embeddings

from urllib.request import urlopen
from PIL import Image
import timm

# get example histology image
img = Image.open(
  urlopen(
    "https://datasets-server.huggingface.co/cached-assets/1aurent/LC25000/--/56a7c495692c27afd294a88b7aadaa7b79d8e270/--/default/train/24999/image/image.jpg"
  )
)

# load model from the hub
model = timm.create_model(
  model_name="hf-hub:1aurent/vit_base_patch16_224.owkin_pancancer_ft_lc25000_colon",
  pretrained=True,
  num_classes=0,
).eval()

# get model specific transforms (normalization, resize)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0))  # output is (batch_size, num_features) shaped tensor

Citation

@article {Filiot2023.07.21.23292757,
  author = {Alexandre Filiot and Ridouane Ghermi and Antoine Olivier and Paul Jacob and Lucas Fidon and Alice Mac Kain and Charlie Saillard and Jean-Baptiste Schiratti},
  title = {Scaling Self-Supervised Learning for Histopathology with Masked Image Modeling},
  elocation-id = {2023.07.21.23292757},
  year = {2023},
  doi = {10.1101/2023.07.21.23292757},
  publisher = {Cold Spring Harbor Laboratory Press},
  URL = {https://www.medrxiv.org/content/early/2023/09/14/2023.07.21.23292757},
  eprint = {https://www.medrxiv.org/content/early/2023/09/14/2023.07.21.23292757.full.pdf},
  journal = {medRxiv}
}