This repository is pytorch implementation of DenseNet.
- python >= 3.8
- pytorch >= 1.4
- torchvision >= 0.5
- pandas >= 1.0
You can train DenseNet by following command.
python run.pyTraining results are put in work/* (* is setting name).
When you want to train DenseNet with different hyper parameters, edit setting.yml and run bellow.
# * means setting name, default is cifar10+BC_k12.
python run.py -S *This repository is MIT-Licensed.
G.Huang, Z.Liu, L.van der Maaten, K.Q.Weinberger. Densely Connected Convolutional Networks. IEEE Conference on Pattern Recognition and Computer Vision (CVPR), 2016.