download_qmnist: Download QMNIST

In jlmelville/snedata: SNE Simulation Dataset Functions

Download QMNIST

Description

Download QMNIST database of handwritten digits. This is the same as the MNIST digits, but with an extended test dataset of 60,000 digits.

Usage

download_qmnist(base_url = qmnist_url, verbose = FALSE)

Arguments

base_url	Base URL that the QMNIST files are located at.
verbose	If TRUE, then download progress will be logged as a message.

Format

A data frame with 785 variables:

px1, px2, px3 ... px784

Integer pixel value, from 0 (white) to 255 (black).

Label

The digit represented by the image, in the range 0-9.

Pixels are organized row-wise. The Label variable is stored as a factor.

There are 120,000 digits in the data set. The first 60,000 are the training set, as found in the qmnist-train-images-idx3-ubyte.hz file. The remaining 60,000 are the test set, from the qmnist-test-images-idx3-ubyte.gz file.

Items in the dataset can be visualized with show_mnist_digit.

For more information see https://github.com/facebookresearch/qmnist.

Details

Downloads the image and label files for the training and test datasets from https://github.com/facebookresearch/qmnist and converts them to a data frame.

Value

Data frame containing the QMNIST digits.

References

Yadav, C., & Bottou, L. (2019, May). Cold Case: The Lost MNIST Digits arXiv preprint arXiv:1905.10498. https://github.com/facebookresearch/qmnist

Examples

## Not run: 
# download the QMNIST data set
qmnist <- download_qmnist()

# first 60,000 instances are the training set
qmnist_train <- head(qmnist, 60000)
# the remaining 60,000 are the test set
qmnist_test <- tail(mnist, 60000)

## End(Not run)

jlmelville/snedata documentation built on Jan. 13, 2024, 2:06 a.m.