The K-nearest neighbors binary classifier of digits 0 and 9 from the SVHN dataset as homework from Deep Learning school. If you want to test it, you should download training and testing datasets and put them in the folder Data inside the project. Links to datasets:
- http://ufldl.stanford.edu/housenumbers/train_32x32.mat
- http://ufldl.stanford.edu/housenumbers/test_32x32.mat
Technology: The model behind DigitsClassifier, consisting of a K-nearest neighbors binary classifier, achieved 70% accuracy with K = 12.