The MNIST_784 dataset is a widely used machine learning benchmark dataset. It consists of 60,000 training and 10,000 test grayscale images of handwritten digits ranging from 0 to 9, making it visually straightforward and easy to work with. Moreover, each image is 28x28 pixels in size, resulting in a total of 784 dimensions when flattened, allowing researchers to explore algorithms in the context of high-dimensional data.

The original train and test images databases from https://yann.lecun.com/exdb/mnist/ have been pre-split into train (60,000 elements) and test (100 elements) sets and stored into a HDF5 file.