-
Notifications
You must be signed in to change notification settings - Fork 5
/
index.html
56 lines (56 loc) · 2.9 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
---
title: DeepLearning.scala
layout: default
type: homepage
---
<!-- Featured -->
<div id="featured">
<div class="container">
<header>
<h2>A simple library for creating complex neural networks</h2>
</header>
<p>
<strong>DeepLearning.scala</strong> is a deep learning toolkit for Scala,<br/>
combining object-oriented and functional programming constructs,<br/>
aims to create statically typed dynamic neural networks<br/>
from <code>map</code>/<code>reduce</code> and other higher order functions.
</p>
<hr />
<div class="row">
<section class="6u" style="height: 26em">
<span class="pennant">
<span style="font-family: Georgia, serif; line-height: 1; font-size: 1.25em">∆</span>
</span>
<h3 style="white-space: nowrap">Differentiable Programming</h3>
<p style="text-align: justify">DeepLearning.scala allows you to build neural networks from mathematical formulas. It supports <a href="https://javadoc.io/page/com.thoughtworks.deeplearning/deeplearning_2.11/latest/com/thoughtworks/deeplearning/plugins/FloatLayers.html">floats</a>, <a href="https://javadoc.io/page/com.thoughtworks.deeplearning/deeplearning_2.11/latest/com/thoughtworks/deeplearning/plugins/DoubleLayers.html">doubles</a>, <a href="https://javadoc.io/page/com.thoughtworks.deeplearning/deeplearning_2.11/latest/com/thoughtworks/deeplearning/plugins/INDArrayLayers.html">GPU-accelerated N-dimensional arrays</a>, and calculates derivatives of the weights in the formulas.</p>
</section>
<section class="6u" style="height: 26em">
<span class="pennant">
<span style="font-family: Georgia, serif; line-height: 1; font-size: 1.25em;">λ</span>
</span>
<h3 style="white-space: nowrap">Functional Programming</h3>
<p style="text-align: justify">
Neural networks are Monads, which can be created by composing higher order functions.
Along with the Monad, we also provide an Applicative type class, to perform multiple calculations in parallel.
</p>
</section>
<section class="6u" style="height: 26em">
<span class="pennant"><span class="fa fa-magic"></span></span>
<h3 style="white-space: nowrap">Dynamic Neural Networks</h3>
<p style="text-align: justify">
Neural networks are programs, too.
All Scala features, including functions, expressions and control flows, are available in neural networks,
which can be even evaluated step by step in a Jupyter Notebook.
</p>
</section>
<section class="6u" style="height: 26em">
<span class="pennant"><span class="fa fa-cogs"></span></span>
<h3 style="white-space: nowrap">Plugins</h3>
<p style="text-align: justify">
DeepLearning.scala supports plugins. There are <a href="/plugins">various plugins</a> providing algorithms, models, hyperparameters or other features.
You can <a href="/get-involved">share your own plugins</a> as simple as creating a Github Gist.
</p>
</section>
</div>
</div>
</div>