Correct code link:<p><a href="https://github.com/chuanli11/MGANs" rel="nofollow">https://github.com/chuanli11/MGANs</a><p>Collection of other implementations of this feedforward neural style transfer approach:<p><a href="https://tensortalk.com/?cat=feedforward-neural-style-transfer" rel="nofollow">https://tensortalk.com/?cat=feedforward-neural-style-transfe...</a><p>Or, regular neural style transfer:<p><a href="https://tensortalk.com/?cat=neural-style-transfer" rel="nofollow">https://tensortalk.com/?cat=neural-style-transfer</a>
They also have a video with some samples and an overview.<p><a href="https://www.youtube.com/watch?v=PRD8LpPvdHI" rel="nofollow">https://www.youtube.com/watch?v=PRD8LpPvdHI</a>
Since it seems this paper's primary focus is on performance, it'd be interesting to see how this technique stacks up against one of those fancy new binary networks (e.g. <a href="http://arxiv.org/abs/1603.05279" rel="nofollow">http://arxiv.org/abs/1603.05279</a>)