Tensorflow implementation of Adversarial Autoencoders

#1

Post by conan7882 » Tue, Aug 28, 2018, 2:03 AM UTC

Tensorflow implementation of Adversarial Autoencoders
github.com

Re: Tensorflow implementation of Adversarial Autoencoders

#2

Post by jingleheimer » Tue, Aug 28, 2018, 5:15 AM UTC

This code is really good.

Re: Tensorflow implementation of Adversarial Autoencoders

#3

Post by polynomial » Tue, Aug 28, 2018, 6:23 AM UTC

For this particular implementation, what are the advantages of using it for supervised or unsupervised learning? (in general?)

And if it's just being used for the MNIST dataset, is there a particular reason for using it in one or the other fashion?

Re: Tensorflow implementation of Adversarial Autoencoders

#4

Post by AstralStorm » Tue, Aug 28, 2018, 8:16 AM UTC

jingleheimer wrote: post #2

This code is really good.

And impossible to use thanks to no license information.

Re: Tensorflow implementation of Adversarial Autoencoders

#5

Post by edtruji » Tue, Aug 28, 2018, 12:22 PM UTC

AstralStorm wrote: post #4

jingleheimer wrote: post #2

This code is really good.

And impossible to use thanks to no license information.

MIT License, it was added after your post https://github.com/conan7882/adversarial-autoencoders-tf/blo...

Re: Tensorflow implementation of Adversarial Autoencoders

#6

Post by amelius » Tue, Aug 28, 2018, 12:56 PM UTC

Question: what framework is used mostly by academic researchers these days for DL?

Re: Tensorflow implementation of Adversarial Autoencoders

#7

Post by nabla9 » Tue, Aug 28, 2018, 1:02 PM UTC

amelius wrote: post #6

Question: what framework is used mostly by academic researchers these days for DL?

Tensorflow, although PyTorch is gaining popularity.

Re: Tensorflow implementation of Adversarial Autoencoders

#8

Post by locuscoeruleus » Tue, Aug 28, 2018, 1:20 PM UTC

polynomial wrote: post #3

For this particular implementation, what are the advantages of using it for supervised or unsupervised learning? (in general?) And if it's just being used for the MNIST dataset, is there a particular reason for using it in one or the other fashion?

From what I can gather the supervised approach allows you to only learn the representation for style rather than which digit it is. The only reason to use one over the other is to demonstrate that it works, I guess?

Re: Tensorflow implementation of Adversarial Autoencoders

#9

Post by amelius » Tue, Aug 28, 2018, 1:34 PM UTC

nabla9 wrote: post #7

amelius wrote: post #6

Question: what framework is used mostly by academic researchers these days for DL?

Tensorflow, although PyTorch is gaining popularity.

Ok. I guess in this case the authors used TF because three of the authors work at Google.

Re: Tensorflow implementation of Adversarial Autoencoders

#10

Post by Lorkki » Tue, Aug 28, 2018, 9:12 PM UTC

jingleheimer wrote: post #2

This code is really good.

Though, in the finest academic tradition, once you try to actually run it, you'll find that it silently depends on a separate library written by the author, which you'll have to find yourself.

(In this case "tensorcv", which is in a separate repository: https://github.com/conan7882/DeepVision-tensorflow )

That aside, I agree that it's an easier read than most ML code.