A deep neural network for MNIST image recognition with the following key features:
- supports unlimited number of layers, nodes and weights (only restriction is memory)
- supports fully connected and convolutional layers
- supports following activation functions: SIGMOID, TANH, RELU
- light weight architecture with a very small memory footprint
- super fast! :-)
- See also
- Deep Neural Network for MNIST Handwriting Recognition
-
http://yann.lecun.com/exdb/mnist/
- Version
- Github Project Page
- Author
- Matt Lind
- Date
- February 2016