Introduction
  • Introduction
  • Computer Vision Problems
  • Linear Layer and Classification Pipeline
  • Loss functions and Softmax
  • Stochastic Gradient Descend
  • PRACTICE #1: Data loading
  • PRACTICE #2: Linear Classifier in PyTorch (part 1)
  • PRACTICE #3: Linear Classifier in PyTorch (part 2)
  • PRACTICE #4: Multi-layer perceptron
Convolutional Neural Networks
  • What is image
  • Motivation to Convolutions
  • Convolution operation
  • Parameters of the convolution
  • Non-linear function
  • Max Pooling and Average Pooling
  • Building deep convolutional network
  • PRACTICE #5: Convolutional Neural Network
Regularization and Normalization
  • Overfitting. L2 regularization
  • DropOut regularization. DropConnect regularization
  • DropBlock regularization
  • Early Stopping regularization
  • Batch Normalization
Improving the quality
  • Data Augmentation
  • Existing datasets
  • Modern Architectures
  • Transfer Learning
Boat Recognition Project
  • Data Loading
  • Data Augmentation
  • Transfer Learning: ResNet-18