News for hackers. Stuff that matters.
Re:Mistakes Programmers Make when Starting in Machine Learning by agibsonccc on Tuesday January 28, 2014 @10:19PM (#7141622)
Binary compatible yes. Star the repo and watch it in the next few days. Example apps are on the way. I plan on implementing a full "easy to use" machine learning lib around this.

Edit: Poke around in the tests, Here's an example of it learning a compressed version of MNIST: https://github.com/agibsonccc/java-deeplearning/blob/master/...

I have a lot more example usage in each of the tests. Test coverage was a higher priority above the documentation, but example usage is there. I'm more than happy to answer emails around the usage of the library as well. I also take feature requests.

I plan on implementing convolutional nets, recursive neural nets and some other ones based around that same structure; That includes the scale out versions with akka for easy multi threading or clustering ( I have built in service discovery with zookeeper among other things in there.

There are no comments on this comment yet.