Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/yasunorikudo/chainer-ResDrop
Deep Networks with Stochastic Depth implementation by Chainer
https://github.com/yasunorikudo/chainer-ResDrop
chainer resdrop
Last synced: 8 days ago
JSON representation
Deep Networks with Stochastic Depth implementation by Chainer
- Host: GitHub
- URL: https://github.com/yasunorikudo/chainer-ResDrop
- Owner: yasunorikudo
- Created: 2016-04-07T09:05:24.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2016-04-11T02:56:22.000Z (over 8 years ago)
- Last Synced: 2024-08-01T22:49:53.319Z (3 months ago)
- Topics: chainer, resdrop
- Language: Python
- Homepage:
- Size: 101 KB
- Stars: 40
- Watchers: 6
- Forks: 8
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- awesome-image-classification - unofficial-chainer : https://github.com/yasunorikudo/chainer-ResDrop
- awesome-image-classification - unofficial-chainer : https://github.com/yasunorikudo/chainer-ResDrop
README
Deep Networks with Stochastic Depth implementation by Chainer
========
Implementation by Chainer. Original paper is [Deep Networks with Stochastic Depth](http://arxiv.org/abs/1603.09382).
This repository includes network definition scripts only.
If you want to train ResDrop from scratch, see [chainer sample code](https://github.com/pfnet/chainer/tree/master/examples/imagenet).
# Requirements
- [Chainer 1.5+](https://github.com/pfnet/chainer) (Neural network framework)
# Usage
In python script, write:
```
from ResDrop152 import ResNet
model = ResNet()
```
# Traning speed
About 25% faster per iteration than ResNet with no layer drop.
# Sample result
I trained ResNet101 with layer drop and ResNet101 with no layer drop for PASCAL VOC Action dataset.
ResNet with layer drop improved the accuracy of test results about 4%.
