Skip to content

Latest commit

 

History

History
25 lines (20 loc) · 1003 Bytes

readme.md

File metadata and controls

25 lines (20 loc) · 1003 Bytes
Raw

JavaCNN

A Java implement of Convolutional Neural Network. Learn from DeepLearnToolbox(https://github.com/rasmusbergpalm/DeepLearnToolbox) more detail. see here(http://www.cnblogs.com/fengfenggirl/p/cnn_implement.html)

Build a CNN

LayerBuilder builder = new LayerBuilder();
builder.addLayer(Layer.buildInputLayer(new Size(28, 28)));
builder.addLayer(Layer.buildConvLayer(6, new Size(5, 5)));
builder.addLayer(Layer.buildSampLayer(new Size(2, 2)));
builder.addLayer(Layer.buildConvLayer(12, new Size(5, 5)));
builder.addLayer(Layer.buildSampLayer(new Size(2, 2)));
builder.addLayer(Layer.buildOutputLayer(10));
CNN cnn = new CNN(builder, 50);

Run on MNIST dataset

String fileName = "data/train.format";
Dataset dataset = Dataset.load(fileName, ",", 784);
cnn.train(dataset, 100);
Dataset testset = Dataset.load("data/test.format", ",", -1);
cnn.predict(testset, "data/test.predict");

It takes a about an hour to complete 100 iteration and get a precison of 97.8%

##Lisence MIT