dlex-cnn : Deep Learning Exercise

A simple convolution neural network framework for exercise.

---

Features

• Lightweight, pure c++ and cuda implemented.
• Easy to understand, and the algorithm implementation details are presented as annotations.

---

Basic Modules

Experiment

Build

Windows: Just open the project file with VS2013 or VS2015, and compile it.

Linux:

Build libdlex_cnn.so

cd linux/cmake
cmake ..
make

Build demo

cd linux/dlex_cnn_test/cmake
cmake ..
make 

License

MIT

Reference

• Caffe: https://github.com/BVLC/caffe
• MxNet: https://github.com/apache/incubator-mxnet
• tiny-dnn: https://github.com/tiny-dnn/tiny-dnn