Milestone 3

  +150

Reorganize my technique for reading MNIST handwriting image data (from the Jan 31st video) into two classes: gray_image and image_set. You can get a sense of the recommended API from gray-image.hh and image-set.hh in the repository. So you should create that file, as well as the corresponding .cc or .cpp files, and test them.

Then, write code to read the label set. These are in separate files from the images. Like the image set, the format is described in the MNIST notes. Again, the recommended API is in label-set.hh, you should use that and create the corresponding .cc or .cpp file, and test.