rhythm92 / chainer-densenet

Densely Connected Convolutional Network implementation by Chainer

Densely Connected Convolutional Network implementation by Chainer

Implementation by Chainer. Original paper is Densely Connected Convolutional Network.

Requirements

• python 2.7

• Chainer 1.15.0+ (Neural network framework)

Start training

For example, run,

python train.py --gpu 0 --batchsize 64 --dataset cifar10 --lr 0.1 --depth 40 --growth_rate 12

Show possible options

python train.py --help

Sample results

• Cifar-10 (batchsize=64, depth=40, growth_rate=12, data augmentation)

Original paper reported 5.24% validation error under the same conditions.

• Cifar-100 (batchsize=64, depth=40, growth_rate=12, data augmentation)

Original paper reported 24.42% validation error under the same conditions.