Fast artistic style transfer by using feed forward network.

• input image size: 512x384
• process time(CPU): 1.954sec (Core i7-5930K)
• process time(GPU): 0.398sec (GPU TitanX)

• Chainer

$ pip install chainer

Download VGG16 model and convert it into smaller file so that we use only the convolutional layers which are 10% of the entire model.

sh setup_model.sh

Need to train one image transformation network model per one style target. According to the paper, the models are trained on the Microsoft COCO dataset.

python train.py -s <style_image_path> -d <training_dataset_path> -g 0

python generate.py <input_image_path> -m <model_path> -o <output_image_path>

This repo has a pretrained model which was trained with "The starry night" by "Vincent van Gogh" as an example.

• example:

python generate.py sample_images/tubingen.jpg -m models/starrynight.model -o sample_images/output.jpg

• Convolution kernel size 4 instead of 3.
• Not sure whether adding/subtracting mean image is needed or not. In this implementation mean image subtraction is done before input image is fed into "image transformation network".

MIT

• Perceptual Losses for Real-Time Style Transfer and Super-Resolution

Codes written in this repository based on following nice works, thanks to the author.

• chainer-gogh Chainer implementation of neural-style. I heavily referenced it.
• chainer-cifar10 Residual block implementation is referred.