The mlpack examples repository contains simple example usages of mlpack. You can take the code here and adapt it into your application, or compile it and see what it does and play with it.
Each of the examples are meant to be as simple as possible, and they are extensively documented.
All the notebooks in this repository can be easily run on https://lab.mlpack.org/.
0. Contents
1. Overview
This repository contains examples of mlpack usage that can be easily adapted to various applications. If you're looking to figure out how to get mlpack working for your machine learning task, this repository would definitely be a good place to look, along with the mlpack documentation.
mlpack is a C++ library that provides machine learning support, but it also provides bindings to other languages, including Python and Julia, and it also provides command-line programs.
Therefore, this repository contains examples not just in C++ but also in other
languages. C++ mlpack usage examples are contained in the c++/ directory;
Python examples in the python/ directory, command-line examples in the
command-line/ directory, and so forth.
2. Building the examples and usage
In order to keep this repository as simple as possible, there is no build
system, and all examples are minimal. For the C++ examples, there is a Makefile
in each example's directory; if you have mlpack installed on your system,
running make should work fine. Some other examples may also use other
libraries, and the Makefile expects those dependencies to also be available.
See the README in each directory for more information, and see the main mlpack
repository and mlpack
website for more information on how to install mlpack.
For Python examples and other-language examples, it's expected that you have mlpack and its dependencies installed.
Each example should be easily runnable and should perform a simple machine learning task on a dataset. You might need to download the dataset first---so be sure to check any README for the example.
3. List of examples
Below is a list of examples available in this repository along with a quick description (just a little bit more than the title):
-
lstm_electricity_consumption: use an LSTM-based recurrent neural network to predict electricity consumption -
lstm_stock_prediction: predict Google's historical stock price (daily high and low) using an LSTM-based recurrent neural network -
spam: predict whether a mobile phone text message in Indonesian is spam or not using logistic regression -
mnist_batch_norm: use batch normalization in a simple feedforward neural network to recognize the MNIST digits -
mnist_cnn: use a convolutional neural network (CNN) similar to LeNet-5 to recognize the MNIST digits -
mnist_simple: use a very simple three-layer feedforward neural network with dropout to recognize the MNIST digits -
mnist_vae_cnn: use a variational autoencoder with convolutional neural networks in the encoder and reparametrization networks to recognize the MNIST digits -
neural_network_regression: use neural network to do regression on Body fat dataset -
q_learning: train a simple deep Q-network agent on CartPole environment
4. Datasets
All the required dataset needed by the examples can be downloaded using the
provided script in the tools directory. You will have to execute
download_data_set.py from the tools/ directory and it will download and
extract all the necessary dataset in order for examples to work perfectly:
cd tools/
./download_data_set.py