My most ambitious project yet, 1,000 points of training data, compressed down into 1,000 list items and one of the largest, if not the largest AI training dataset to ever be uploaded to the scratch website. MiniNIST Is a small Dataset for the training of a Handwritten Digit Recognition Neural Network (HDRNN), it has 1,000 points of data, 100 examples of digits from 0-9 each, in a 24x24 grid, compressed heavily with a simple algorithm (Decompression algorithm provided and demonstrated in project). MicroNIST was the original MiniNIST and is a test data set of 100 examples, 10 examples for each digit. The dataset was originally presented as raw data, however I decided to decrease the scratch project file size by also running it through the Run Length Encoding (RLE) compression algorithm.
