forked from SciSharp/NumSharp
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathNDArray.Creation.html
More file actions
178 lines (155 loc) · 7.16 KB
/
NDArray.Creation.html
File metadata and controls
178 lines (155 loc) · 7.16 KB
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
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
<!DOCTYPE html>
<!--[if IE]><![endif]-->
<html>
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1">
<title>Array creation </title>
<meta name="viewport" content="width=device-width">
<meta name="title" content="Array creation ">
<meta name="generator" content="docfx 2.40.2.0">
<link rel="shortcut icon" href="../favicon.ico">
<link rel="stylesheet" href="../styles/docfx.vendor.css">
<link rel="stylesheet" href="../styles/docfx.css">
<link rel="stylesheet" href="../styles/main.css">
<meta property="docfx:navrel" content="../toc.html">
<meta property="docfx:tocrel" content="toc.html">
</head>
<body data-spy="scroll" data-target="#affix" data-offset="120">
<div id="wrapper">
<header>
<nav id="autocollapse" class="navbar navbar-inverse ng-scope" role="navigation">
<div class="container">
<div class="navbar-header">
<button type="button" class="navbar-toggle" data-toggle="collapse" data-target="#navbar">
<span class="sr-only">Toggle navigation</span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
<span class="icon-bar"></span>
</button>
<a class="navbar-brand" href="../index.html">
<img id="logo" class="svg" src="../logo.svg" alt="">
</a>
</div>
<div class="collapse navbar-collapse" id="navbar">
<form class="navbar-form navbar-right" role="search" id="search">
<div class="form-group">
<input type="text" class="form-control" id="search-query" placeholder="Search" autocomplete="off">
</div>
</form>
</div>
</div>
</nav>
<div class="subnav navbar navbar-default">
<div class="container hide-when-search" id="breadcrumb">
<ul class="breadcrumb">
<li></li>
</ul>
</div>
</div>
</header>
<div role="main" class="container body-content hide-when-search">
<div class="sidenav hide-when-search">
<a class="btn toc-toggle collapse" data-toggle="collapse" href="#sidetoggle" aria-expanded="false" aria-controls="sidetoggle">Show / Hide Table of Contents</a>
<div class="sidetoggle collapse" id="sidetoggle">
<div id="sidetoc"></div>
</div>
</div>
<div class="article row grid-right">
<div class="col-md-10">
<article class="content wrap" id="_content" data-uid="">
<h1 id="array-creation">Array creation</h1>
<p>Before we do some fancy numeric stuff or even machine learning we have to clear one thing.</p>
<p><strong>How do we generate NDArrays?</strong></p>
<p>Since NDArray is the key class in SciSharp stack there must be numerous possibilities how to generate this arrays. And yes that’s the case.</p>
<p>Maybe first of all we should see the dump way – which can be always used but is not too user friendly.<br>
In this example we access the Storage property directly - one more reason to avoid it.</p>
<p><strong>Dump way</strong></p>
<pre><code class="lang-CSHARP">// first constructor with data type and shape (here 3x3 matrix)
var nd = new NDArray(typeof(double),3,3);
// set 9 elements into the storage of this array.
np.Storage.SetData(new double[] {1,2,3,4,5,6,7,8,9});
</code></pre>
<p>Ok looks not too difficult. But also not too userfriendly.</p>
<p>We create an empty NDArray with 3x3 shape, fill it with 9 elements.
We followed the row wise matrix layour by default.</p>
<p>So with this 3x3 shaped NDArray we can do matrix multiplication, QR decomposition, SVD, ...</p>
<p>But keep in mind - always be careful with your shape and be sure what you want to do with your elements in this shape.</p>
<p><strong>Create by enumeration</strong></p>
<p>The next example shows the numpy style creation.</p>
<pre><code class="lang-CSHARP">using NumSharp.Core;
// we take the Data / elements from an array
// we do not need to define the shape here - it is automaticly shaped to 1D
var np1 = np.array(new double[] {1,2,3,4,5,6} );
</code></pre>
<p>Ok as we can see, this time the array was created without define the shape.</p>
<p>This is possible since the method expect that the double[] array shall be transformed into a NDArray directly.</p>
<p><strong>Create by implicit cast</strong></p>
<p>Beside this numpy style C# offers its own flavour of creation.</p>
<pre><code class="lang-CSHARP">using NumSharp.Core;
// implicit cast double[] to NDArray - dtype & shape are deduced by array type and shape
NDArray nd = new double[]{1,2,3,4};
</code></pre>
<p>And for matrix and n dim tensors also work the same.</p>
<pre><code class="lang-CSHARP">using NumSharp.Core;
NDArray nd = new double[,]{{1,2,3},{4,5,6}};
</code></pre>
<p>Beside the .NET array to NDArray converting there exist different kinds of methods which also exist in numpy.</p>
<p><strong>Create by given range</strong></p>
<pre><code class="lang-CSHARP">using NumSharp.Core;
// we simple say "create an array with 10 elements"
// start is expected to be 0 and step 1
var np1 = np.arange(10);
// this time start with 1, step 2
// and do it as long as smaller than 10
var np2 = np.arange(1,10,2);
</code></pre>
<p><strong>Create diagonal matrix</strong></p>
<pre><code class="lang-CSHARP">using NumSharp.Core;
// simple 5x5 eye matrix
var nd1 = np.eye(5);
// 3x3 eye matrix but elements different diagonal
nd1 = np.eye(3,1);
</code></pre>
<p><strong>Create by linspace</strong></p>
<pre><code class="lang-CSHARP">using NumSharp.Core;
// create vector with 50 elements, from 4 to 10
// include last element
// and convert them to double (float64)
var nd1 = np.linspace(4,10, 50, true, np.float64);
</code></pre>
</article>
</div>
<div class="hidden-sm col-md-2" role="complementary">
<div class="sideaffix">
<div class="contribution">
<ul class="nav">
<li>
<a href="https://github.com/dotChris90/NumSharp/blob/master/docfx_project/articles/NDArray.Creation.md/#L1" class="contribution-link">Improve this Doc</a>
</li>
</ul>
</div>
<nav class="bs-docs-sidebar hidden-print hidden-xs hidden-sm affix" id="affix">
<!-- <p><a class="back-to-top" href="#top">Back to top</a><p> -->
</nav>
</div>
</div>
</div>
</div>
<footer>
<div class="grad-bottom"></div>
<div class="footer">
<div class="container">
<span class="pull-right">
<a href="#top">Back to top</a>
</span>
<span>Generated by <strong>DocFX</strong></span>
</div>
</div>
</footer>
</div>
<script type="text/javascript" src="../styles/docfx.vendor.js"></script>
<script type="text/javascript" src="../styles/docfx.js"></script>
<script type="text/javascript" src="../styles/main.js"></script>
</body>
</html>