top | item 10991544

(no title)

andreaferretti | 10 years ago

I think Nim would be ideal for that kind of applications, in particular neural networks stuff. But definitely we are not there yet. I try to explain my reasons in this blog post: http://rnduja.github.io/2015/10/21/scientific-nim/

To get to a point where we have something good to use, though, there are a few libraries to develop. I started working out a basic layer for linear algebra, that is needed in many applications - see http://unicredit.github.io/linear-algebra/

I hoped I would be able to start working on some neural network stuff based on that, but I did not find the time yet...

discuss

order

Lovenim|10 years ago

Interesting post. I also enjoyed the posts about Torch but when installing in ubuntu 15.10 there were some problems. Libraries are necessary to avoid reinventing the wheel.