README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: label
      dtype:
        class_label:
          names:
            '0': benign
            '1': malignant
            '2': normal
  splits:
    - name: train
      num_bytes: 211946420
      num_examples: 624
    - name: test
      num_bytes: 53658067
      num_examples: 156
  download_size: 265190925
  dataset_size: 265604487
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
task_categories:
  - image-classification
language:
  - en
size_categories:
  - n<1K

Breast Cancer Ultrasound Images

This dataset has been created by hand for scientific and learning purposes. It is used in "Fine-tuning a Vision Transformer Model With a Custom Biomedical Dataset" in Hugging Face Cookbook.

Dataset Details

The dataset contains total of 780 images from the following 3 classes: benign, malignant and normal. You can use this dataset for your computer vision tasks.

Used model

You can use this fine-tuned Vision Transformer Model with this dataset: emre570/google-vit-large-finetuned