Deep Declarative Networks (DDNs) are a class of deep learning model that allows for optimization problems to be embedded within an end-to-end learnable network. This repository maintains code, tutorials and other resources for developing and understanding DDN models.
You can find more details in this paper (also here), which if you would like to reference in your research please cite as:
@techreport{Gould:PrePrint2019,
author = {Stephen Gould and
Richard Hartley and
Dylan Campbell},
title = {Deep Declarative Networks: A New Hope},
eprint = {arXiv:1909.04866},
institution = {Australian National University (arXiv:1909.04866)},
month = {Sep},
year = {2019}
}
or
@journal{Gould:PAMI2022,
author = {Stephen Gould and
Richard Hartley and
Dylan Campbell},
title = {Deep Declarative Networks},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
year = {2022},
month = {Aug},
volume = {44},
pages = {3988--4004},
doi = {10.1109/TPAMI.2021.3059462}
}
When running code from the command line make sure you add the ddn
package to your PYTHONPATH. For example:
export PYTHONPATH=${PYTHONPATH}:ddn
python tests/testBasicDeclNodes.py
Documentation for the ddn
package is provided in the ddn
directory and many examples given in the interactive turotials
.
These should be opened in Jupyter notebook:
cd tutorials
jupyter notebook
or viewed using using jupyter.org's notebook viewer.
Reference (PyTorch) applications for image and point cloud classification can be found under the apps
directory. See the README
files therein for instructions on installation and how to run.
The ddn
library is distributed under the MIT license. See the LICENSE file for details.