BigDEAL Forecasting Competition 2018 - LSTM example

Code for training LSTM Neural Networks for energy load forecasting. The competition was hosted by Dr. Tao Hong in Novemeber of 2018 : http://blog.drhongtao.com/2018/11/leaderboard-for-bfcom2018-qualifying-match.html

Defining Models

• Define models in modelLib.py

• Can hard code parameters in or make them parameteric and set them inside config.toml

Training Models

• run train.py after setting the values in config.toml to your liking

• models will be saved inside your model directory modelDir

• the config.toml file used in training will also be copied to the model directory

Testing Models

• run 'test.py' with the path to the config.toml file of the model as the first argument

  python test.py ./models/LSTM01_001/config.toml