Chainer Tutorial

• Introduction to Chainer
  • Core Concept
  • Forward/Backward Computation
  • Links
  • Write a model as a chain
  • Optimizer
  • Trainer
  • Serializer
  • Example: Multi-layer Perceptron on MNIST
• How to Write a New Network
  • Convolutional Network for Visual Recognition Tasks
  • Recurrent Nets and their Computational Graph
• Using GPU(s) in Chainer
  • Relationship between Chainer and CuPy
  • Basics of cupy.ndarray
  • Run Neural Networks on a Single GPU
  • Model-parallel Computation on Multiple GPUs
  • Data-parallel Computation on Multiple GPUs with Trainer
  • Data-parallel Computation on Multiple GPUs without Trainer
• Define your own function
  • Differentiable Functions
  • Unified forward/backward methods with NumPy/CuPy functions
  • Write an Elementwise Kernel Function
  • Links that wrap functions
  • Testing Function
• Type check
  • Basic usage of type check
  • Detail of type information
  • Internal mechanism of type check
  • More powerful methods
  • Call functions
  • More complicated cases
  • Typical type check example