Tensorflow implementation of CondenseNet: An Efficient DenseNet using Learned Group Convolutions. The code is tested with cifar10, inference phase not implemented yet.

Official PyTorch implementation by @ShichenLiu here.

• Python 2.7+ (3.5+ is recommended)
• NumPy
• TensorFlow 1.0+

• CIFAR-10

• Go to data/ folder and run python2 generate_cifar10_tfrecords.py --data-dir=./cifar-10-data. This code is directly borrowed from tensorflow official repo and have to be run with python 2.7+.

Use default parameters:

python main.py

Check out tunable hyper-parameters:

python main.py --help

Other parameters including stages, groups, condense factor, and growth rate are in experiment.py.

• Training for 300 epochs with the default settings reach testing accuracy 93.389% (paper report is 94.94%). There might be some details I didn't notice, feel free to point them out.
• All the default parameters settings follow the paper/official pytorch implementation.
• Current implmentations of standard group convolution and learned group convolution are very inefficient (a bunch of reshape, transpose and concat), looking for help to build much more efficient graph.
• Evaluation phase (index select) has not been implemented yet, looking for potential help as well :D.
• Issues are welcome!

• The paper.
• Official PyTorch Implementation.