In this module are implemented some generic `DataBox`..

image_box[source]

image_box(size=224, folder='', aug_kwargs=None, **kwargs)

Make a DataBox from an ImageBlock.

Parameters
----------
size: int, default=224
    determine to which size images should be resized.
folder: path-like, optional
    select only images from `folder` path.
aug_kwargs: dict, default={}
    add basic data augmentation to the training loop with the
    albumentation library ; for more information look at
    aug_transforms function documentation in fastai.
kwargs
    additionnal keyword arguments for DataBlok interface.

get_array[source]

get_array(suffix)

Return a getter that loads a numpy array with given suffix.

cat_box[source]

cat_box(suffix='')

Make a Category Box from the categorial array of the dataset.

mcat_box[source]

mcat_box(suffix='')

Make a Category Box from the categorial array of the dataset.

bbox_box[source]

bbox_box()