- class torchvision.datasets.PCAM(root: str, split: str = 'train', transform: Optional[Callable] = None, target_transform: Optional[Callable] = None, download: bool = False)¶
The PatchCamelyon dataset is a binary classification dataset with 327,680 color images (96px x 96px), extracted from histopathologic scans of lymph node sections. Each image is annotated with a binary label indicating presence of metastatic tissue.
This dataset requires the
h5pypackage which you can install with
pip install h5py.
root (string) – Root directory of the dataset.
split (string, optional) – The dataset split, supports
transform (callable, optional) – A function/transform that takes in a PIL image and returns a transformed version. E.g,
target_transform (callable, optional) – A function/transform that takes in the target and transforms it.
download (bool, optional) – If True, downloads the dataset from the internet and puts it into
root/pcam. If dataset is already downloaded, it is not downloaded again.