Densely Connected Convolutional Networks (DenseNets)

A Tensorflow 2.0 reimplementation of DenseNet.

Citation

Densely Connected Convolutional Networks won Best Paper at CVPR 2017.

@inproceedings{huang2017densely,
    title={Densely connected convolutional networks},
    author={Huang, Gao and Liu, Zhuang and van der Maaten, Laurens and Weinberger, Kilian Q },
    booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
    year={2017}
}

Requirements

This implementation assumes python==3.7.3 and tensorflow-gpu==2.0.0. All requirements are listed in requirements.txt. If you wish to run without gpu support, specify tensorflow==2.0.0 in requirements.txt before continuing. To install, run:

pip install -r requirements.txt

Formatting and Linting

We use black, flake8, mypy, and isort to format and lint code. Before contributing, make sure these packages are installed:

pip install black flake8 mypy isort

Then run the following:

# fix formatting
black .

# check for errors
flake8 --config=.flake8

# fix import sort order
isort

# type checking
mypy --config-file mypy.ini code

Formatting configuration files should automatically be picked up if you are using an editor that supports these plugins, e.g., vscode.