EMNIST

Extended MNIST - Python Package

The EMNIST Dataset

The EMNIST Dataset is an extension to the original MNIST dataset to also include letters. For more details, see the EMNIST web page and the paper associated with its release:

Cohen, G., Afshar, S., Tapson, J., & van Schaik, A. (2017). EMNIST: an extension of MNIST to handwritten letters. Retrieved from http://arxiv.org/abs/1702.05373

The EMNIST Python Package

This package is a convenience wrapper around the EMNIST Dataset. The package provides functionality to automatically download and cache the dataset, and to load it as numpy arrays, minimizing the boilerplate necessary to make use of the dataset. (NOTE: The author of the Python package is not affiliated in any way with the authors of the dataset and the associated paper.)

Installation

To install the EMNIST Python package along with its dependencies, run the following command:

pip install emnist

For optional support for visual inspection of the data via emnist.inspect(), run:

pip install emnist[inspect]

The dataset itself is automatically downloaded and cached when needed. To preemptively download the data and avoid a delay later during the execution of your program, execute the following command after installation:

python -c "import emnist; emnist.ensure_cached_data()"

Alternately, if you have already downloaded the original IDX-formatted dataset from the EMNIST web page, copy or move it to ~/.cache/emnist/, where ~ is your home folder, and rename it from gzip.zip to emnist.zip. The package will use the existing file rather than downloading it again.

Usage

Usage of the EMNIST Python package is designed to be very simple.

To get a listing of the available subsets:

>>> from emnist import list_datasets
>>> list_datasets()
['balanced', 'byclass', 'bymerge', 'digits', 'letters', 'mnist']

(See the EMNIST web page for details on each of these subsets.)

To load the training samples for the 'digits' subset:

>>> from emnist import extract_training_samples
>>> images, labels = extract_training_samples('digits')
>>> images.shape
(240000, 28, 28)
>>> labels.shape
(240000,)

To load the test samples for the 'digits' subset:

>>> from emnist import extract_test_samples
>>> images, labels = extract_test_samples('digits')
>>> images.shape
(40000, 28, 28)
>>> labels.shape
(40000,)

Data is extracted directly from the downloaded compressed file to minimize disk usage, and is returned as standard numpy arrays.

To visually inspect samples from the data set, you will need to ensure you have additional requirements installed via pip install emnist[inspect].

>>> from emnist import inspect
>>> inspect('digits')