Don Tsang

@Kycklingar it is also used in http://endchan.xyz/ so you know it is pretty good

Don't forget QWERTZ (Germany, Austria and Eastern Europe)!

@dmuth please release your 6 dice or 7 dice wordlist, it would be very useful for those who are curious

@dmuth I meant the actual data and not the script that generates the data (for the sale of replication)

Agreed, but the problem is that more and more papers are coming out, especially with the recorrection with https://github.com/nightrome/really-awesome-gan Is it possible to judge a paper based on the amount...

In that case, besides popularity, there should be a "uniqueness" factor, since many GANs can be abstracted away to having the same meta-structure. See https://github.com/kozistr/Awesome-GANs/issues/18 for some examples. Of course...

I would actually recommend Tensorflow 2 as it is the most popular, and that it has a distributed network training model that may be useful in some way.

Just checked the popularity of Tensorflow 2 over PyTorch, and they are on par with one another... Are there any Keras-like libraries that support both TF and PyTorch? (I guess...

So I would make that claim that whatever is more popular and easier to implement the underlying features should be used, that would be Tensorflow in this case.

I would like to make the note that visual descriptions are somewhat tangentially related to GAN structure but not completely, as GANs with same underlying structure may have different names...