CS231n-2019

Notes and assignment solution to http://cs231n.stanford.edu

Assignment 1

• k-Nearest Neighbor (kNN) solution
  • Notebook, k_nearest_neighbor
• Training a Support Vector Machine solution
  • Notebook, linear_svm, linear_classifier
• Implement a Softmax classifier solution
  • Notebook, softmax
• Two-Layer Neural Network solution
  • Notebook, neural_net
• Higher Level Representations: Image Features solution
  • Notebook

Assignment 2

• Fully-connected Neural Network
  • Notebook, optim
• Batch Normalization
  • Notebook
• Dropout
  • Notebook, fc_net
• Convolutional Networks
  • Notebook, cnn, layers
• PyTorch on CIFAR-10
  • Notebook

Assignment 3

• Image Captioning with Vanilla RNNs
  • Notebook
• Image Captioning with LSTMs
  • Notebook, rnn_layers, rnn

Notes

• Hinge loss derivative
• Softmax loss derivative
• Computational graph 2-layer-NN
• Batchnorm computational graph
• Batchnorm alt