https://github.com/yasunorikudo/chainer-DenseNet

Densely Connected Convolutional Network implementation by Chainer
https://github.com/yasunorikudo/chainer-DenseNet

chainer densenet

Last synced: about 2 months ago
JSON representation

Densely Connected Convolutional Network implementation by Chainer

• Host: GitHub
• URL: https://github.com/yasunorikudo/chainer-DenseNet
• Owner: yasunorikudo
• Created: 2016-09-03T09:37:33.000Z (about 8 years ago)
• Default Branch: master
• Last Pushed: 2017-07-15T10:50:27.000Z (about 7 years ago)
• Last Synced: 2024-04-07T16:34:19.355Z (6 months ago)
• Topics: chainer, densenet
• Language: Python
• Homepage:
• Size: 563 KB
• Stars: 39
• Watchers: 4
• Forks: 10
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-very-deep-learning - Chainer Implementation

README

        
Densely Connected Convolutional Network implementation by Chainer

========

Implementation by Chainer. Original paper is [Densely Connected Convolutional Network](https://arxiv.org/abs/1608.06993).

# Requirements

- python 2.7

- [Chainer 1.19.0+](https://github.com/pfnet/chainer) (Neural network framework)

# Start training

For example, run,

```

python train.py --gpus 0 --batchsize 64 --dataset cifar10 --lr 0.1 --depth 100 --growth_rate 24 --split_size 4

```

## Show possible options

```

python train.py --help

```

# Sample results

- Cifar-10 (batchsize=64, depth=100, growth_rate=24, with data augmentation)

![](https://raw.githubusercontent.com/yasunorikudo/chainer-DenseNet/images/cifar10.png)

[Original paper](https://arxiv.org/abs/1608.06993) reported 3.74% validation error under the same conditions.

- Cifar-100 (batchsize=64, depth=100, growth_rate=24, with data augmentation)

![](https://raw.githubusercontent.com/yasunorikudo/chainer-DenseNet/images/cifar100.png)

[Original paper](https://arxiv.org/abs/1608.06993) reported 19.25% validation error under the same conditions.