This is a simple linear classifier using SVM on dataset CIFAR-10. Most of the code is based on an assignment of CS231n.
-
linear_classifier.ipynb: it's the main file and run on jupyter notebook so that you can interactively execute each part of code.
-
data_utils.py: this file contains functions to load the dataset in code.
-
gradient_check.py: it do gradient check, i.e, it compares the gradient computed analytically with the numerically method.
-
classifiers/linear_svm.py: here we calculate the SVM loss and gradient.
-
classifiers/linear_classifier.py: it's the model itself, where we can train and predict.