Deep Residual Network implementation by Chainer

Implementation by Chainer. Original paper is Deep Residual Learning for Image Recognition.

This repository includes network definition scripts, caffemodel converter and prediction scripts.

If you want to train ResNet from scratch, see chainer sample code

Requirements

• Chainer 1.5+ (Neural network framework)

Preparation

If you do not use Caffe, jump to "Download Chainer Model"

Caffe SetUp

At chainer-ResNet root folder

Run:

git clone --recursive https://github.com/KaimingHe/deep-residual-networks.git
cd deep-residual-networks/caffe  

After making Makefile.config

Run:

make -j16 && make pycaffe

Convert Caffe Model to Chainer Model

Download Caffe models here. Place Caffe models at $ROOT/data

Run:

python load_model.py --model ResNet152

Download Chainer Model

If you do not want to use Caffe, you can download Chainer models.

Run:

wget https://www.dropbox.com/s/4mvqcp9qr9ldjgc/ResNet50.model
wget https://www.dropbox.com/s/yqasroj1poru24u/ResNet101.model
wget https://www.dropbox.com/s/he2ndi7f9jixak6/ResNet152.model

Predict

Run:

python predict.py --model ResNet152 --img cat.png