My solutions for cs231n standford course assignments.
The assignments include implementations of Machine Learning algorithms like KNN, Softmax Classification, Multiclass SVM, Fully Connected Neural Networks, Concolutional Neural Networks to Image Classification on CIFAR10 dataset. In addition to implementations of algorithms they also involve stuff like Hyperparameter Tuning, Debugging, Testing and Visualizing weights.
An interesting takeaway from these assignments is learning how to write efficient vectorized implementaions of Machine Learning algorithms using Python, Numpy. Using the vectorized implementations the runtime of even the most trivial algorithm like KNN can be improved over 100 times, compared to the regular for loop implementations.