This repository is my solution to the homeworks given in the MICHIGAN EECS 498.008 / 598.008 - Deep Learning for Computer Vision course.
The homework included in the repo:
- PyTorch101 & KNN (k-nearest neighbors)
- Linear Classifier (SVM loss, SoftMax loss) & Two-Layer NN
The learned weights of each class of the CIFAR10 dataset, using an SVM linear classifier:
Train and Validation accuracies of a two layer NN, for differnet hidden layer sizes (The network architecture is Input-FC-ReLU-FC-SoftMax):
