/DNN-NILM-template

The current repository offer a ready to use project for easy design of experiments based on

Primary LanguagePythonMIT LicenseMIT

DNN-NILM-experiment

The current repository is template-project for DNN-NILM scholars. The project is compatible with the latest version of NILMtk as well as the deep-nilmtk. It is aimed to offer a more inclusive environment/toolkit that can be used by PyTorch community as well as the Tensorflow community. The goal is to decouple the NILM development from the deep learning framework as well to improve the efficiency of research.

The structure of the template

The repository suggests a pre-configured project for experimenting with deep-nilmtk and benefiting from the

  1. requirements.txt: A file containing python requirements for executing of the experiment.
  2. src: This folder is expected to contain the source code for the experimental setup. It has the following structure:
  • model: A folder containing the code of the model.
  • data: A containing python scripts for data pre-processing, data loaders and data post-processing.
  • experiment.py: A python script setting up the code for the experiment.
  1. dataset: This repository is expected to contain the NILM datasets, in the hdf5 format, that will be used during the experimental setup.

How to run ?

docker build -t experiment_setup .
docker run --gpus 'all' --name exp_exec experiment_setup
docker cp exp_exec:/home/guestuser/model_evaluation ./results

Author

Hafsa Bousbiat, email: hafsa.bousbiat@gmail.com

Copyright and license

Code released under the MIT Licence.