neuralstyletransfer

This is the python library which takes two image content image and style image and provide the output image looks like a content image but painted in the style of provided reference style image

Usage

pip install neural-style-transfer

Import NeuralStyleTransfer class from Library

from neuralstyletransfer.style_transfer import NeuralStyleTransfer

create class object

nst = NeuralStyleTransfer()

Load the content image and style image from local path

nst.LoadContentImage(content_img_path)
nst.LoadStyleImage(style_img_path)

You can also load image from external URL by providing pathType='url' by default it is setted to 'local'

nst.LoadContentImage(content_url, pathType='url')
nst.LoadStyleImage(style_url, pathType='url')

Use default parameters to train the model

Default parameters are given below we can change it based on requirement by providing values on calling .apply method

• contentWeight=1e3
• styleWeight=1e-2
• epochs=500

output = nst.apply()
output.save('output.jpg')

Content image

Style image

Output image