TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Show HN: Fast Style Transfer in TensorFlow

184 pointsby loganeover 8 years ago

10 comments

minimaxirover 8 years ago
Terminology note: there is a difference between Style Transfer and <i>Fast</i> Style Transfer.<p>Fast Style Transfer takes a <i>very</i> long time to train a style model (and the output models can be somewhat large; the pre-trained models in the Dropbox are 20MB), but the styles can be applied quickly. Fast Style Transfer is the technique that is used by Prisma&#x2F;Facebook. This repo is the first I&#x27;ve seen that uses TensorFlow instead of lua&#x2F;Torch dependency shenanigans, and as a result should be much easier to set up. (this code release also beats Google&#x27;s TF code release for their implementation: <a href="https:&#x2F;&#x2F;research.googleblog.com&#x2F;2016&#x2F;10&#x2F;supercharging-style-transfer.html" rel="nofollow">https:&#x2F;&#x2F;research.googleblog.com&#x2F;2016&#x2F;10&#x2F;supercharging-style-...</a> )<p>EDIT: Playing around with this repository, I can stylize a 1200x1200px image on a dual-core CPU on a 2013 rMBP in about 30 seconds.<p>Normal Style Transfer is ad-hoc between the style image and the initial image; it can only be used for one image at a time, but overall it is faster than training a model for Fast Style Transfer (however, this is infeasible for mobile devices).
评论 #12843220 未加载
dharma1over 8 years ago
Nice work! Have you seen this - could help reduce some of those checkerboard artifacts<p><a href="http:&#x2F;&#x2F;distill.pub&#x2F;2016&#x2F;deconv-checkerboard&#x2F;" rel="nofollow">http:&#x2F;&#x2F;distill.pub&#x2F;2016&#x2F;deconv-checkerboard&#x2F;</a>
评论 #12839686 未加载
评论 #12839157 未加载
hacker_9over 8 years ago
Utterly amazing what people are achieving with neural nets. The idea that &#x27;style transfer&#x27; can be fit into an algorithm is slightly blowing my mind right now.<p>The jumping fox video does looks a bit &#x27;off&#x27; though, I think because the animation is kept the same and so it ends up looking too realistic for that style. Still this is early days!
评论 #12839918 未加载
santaclausover 8 years ago
There has been a lot of cool image manipulation&#x2F;transformation&#x2F;synthesis work coming out over the past couple of years using NNs and such. I&#x27;m curious if any of these techniques have started worming their way into products? Will new effects in Photoshop (or whatever) get better over point releases as companies train up better and better NNs?
lhnzover 8 years ago
It would be nice if this algorithm could be created in a more localised form, so that you could take a photo and apply the effect with different intensities as if you were using a brush.<p>Merely because sometimes it gets things wrong, and I think a normal human could correct things if they had more control over the process.
评论 #12846241 未加载
评论 #12846170 未加载
eitallyover 8 years ago
Relevant Tensorflow&#x2F;Brain blog post: <a href="https:&#x2F;&#x2F;research.googleblog.com&#x2F;2016&#x2F;10&#x2F;supercharging-style-transfer.html" rel="nofollow">https:&#x2F;&#x2F;research.googleblog.com&#x2F;2016&#x2F;10&#x2F;supercharging-style-...</a>
pizzaover 8 years ago
So you can transfer style from image A to image B. What I want to know is, can you use style transfer to &quot;amplify&quot; image A&#x27;s style to, say, 500%?
评论 #12841065 未加载
spacehackerover 8 years ago
Nice work. I&#x27;d be curious how this technique compares with other fast style transfer methods, if any exist.
visargaover 8 years ago
Results look pretty nice. Congrats!
rubyfanover 8 years ago
This is ridiculously cool.