• How-Tos
  • construct the network model
    • sequential model
    • graph model
  • train the model
    • regression
    • classification
      • train graph model
    • “freeze” layers
  • use/evaluate trained model
    • predict a value
    • evaluate accuracy
    • calculate the loss
  • visualize the model
    • visualize graph networks
    • visualize each layer activations
    • visualize convolution kernels
  • io
    • save and load the model
    • import caffe’s model
    • reading data
    • reading images
  • get/set the properties
    • traverse layers
    • get layer types
    • get weight vector
    • change the weight initialization
    • change the seed value
  • tune the performance
    • profile
    • change the number of threads while training
  • handle errors
    • catch application exceptions
• Integrate with your application
  • Step1/3: Include tiny_dnn.h in your application
  • Step2/3: Enable C++11 options
    • Visual Studio(2013-)
    • gcc(4.8-)/clang(3.3-)
  • Step3/3: Add include path of tiny-dnn to your build system
• Train network with your original dataset
  • 1. using opencv (image file => vec_t)
  • 2. using mnisten (image file => idx format)
  • 3. from levelDB (caffe style => [vec_t, label_t])