Machine Learning with Swift

tiny-dnn

tiny-dnn is a header-only convolutional neural network framework written in C++.

The following is an example of neural network creation from the official documentation:

network<sequential> net; 
// add layers 
net << conv<tan_h>(32, 32, 5, 1, 6)  // in:32x32x1, 5x5conv, 6fmaps 
    << ave_pool<tan_h>(28, 28, 6, 2) // in:28x28x6, 2x2pooling 
    << fc<tan_h>(14 * 14 * 6, 120)   // in:14x14x6, out:120 
    << fc<identity>(120, 10);        // in:120,     out:10 

You can train your models or you can convert pre-trained Caffe models using the caffe_converter tool. It supports a handful of acceleration types. OpenCV is a dependency for iOS. You can find an iOS example on tiny-dnn's GitHub repository.

GitHub repository: https://github.com/tiny-dnn/tiny-dnn