Deci-AI
/
super-gradients
connected to https://github.com/Deci-AI/super-gradients.git


  
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

	
179

	
180

	
181

	
182

	
183

	
184

	
185

	
186

	
187

	
188

	
189

	
190

	
191

	
192

	
193

	
194

	
195

	
196

	
197

	
198

	
199

	
200

	
201

	
202

	
203

	
204

	
205

	
206

	
207

	
208

	
209

	
210

	
211

	
212

	
213

	
214

	
215

	
216

	
217

	
218

	
219

	
220

	
221

	
222

	
223

	
224

	
225

	
226

	
227

	
228

	
229

	
230

	
231

	
232

	
233

	
234

	
235

	
236

	
237

	
238

	
239

	
240

	
241

	
242

	
243

	
244

	
245

	
246

	
247

	
248

	
249

	
250

	
251

	
252

	
253

	
254

	
255

	
256

	
257

	
258

	
259

	
260

	
261

	
262

	
263

	
264

	
265

	
266

	
267

	
268

	
269

	
270

	
271

	
272

	
273

	
274

	
275

	
276

	
277

	
278

	
279

	
280

	
281

	
282

	
283

	
284

	
285

	
286

	
287

	
288

	
289

	
290

	
291

	
292

	
293

	
294

	
295

	
296

	
297

	
298

	
299

	
300

	
301

	
302

	
303

	
304

	
305

	
306

	
307

	
308

	
309

	
310

	
311

	
312

	
313

	
314

	
315

	
316

	
317

	
318

	
319

	
320

	
321

	
322

	
323

	
324

	
325

	
326

	
327

	
328

	
329

	
330

	
331

	
332

	
333

	
334

	
335

	
336

	
337

	
338

	
339

	
340

	
341

	
342

	
343

	
344

	
345

	
346

	
347

	
348

	
349

	
350

	
351

	
352

	
353

	
354

	
355

	
356

	
357

	
358

	
359

	
360

	
361

	
362

	
363

	
364

	
365

	
366

	
367

	
368

	
369

	
370

	
371

	
372

	
373

	
374

	
375

	
376

	
377

	
378

	
379

	
380

	
381

	
382

	
383

	
384

	
385

	
386

	
387

	
388

	
389

	
390

	
391

	
392

	
393

	
394

	
395

	
396

	
397

	
398

	
399

	
400

	
401

	
402

	
403

	
            <!DOCTYPE html>
<html class="writer-html5" lang="en" >
<head>
  <meta charset="utf-8" />
  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
  <title>super_gradients.training.datasets.mixup &mdash; SuperGradients 1.0 documentation</title>
      <link rel="stylesheet" href="../../../../_static/pygments.css" type="text/css" />
      <link rel="stylesheet" href="../../../../_static/css/theme.css" type="text/css" />
      <link rel="stylesheet" href="../../../../_static/graphviz.css" type="text/css" />
  <!--[if lt IE 9]>
    <script src="../../../../_static/js/html5shiv.min.js"></script>
  <![endif]-->
  
        <script data-url_root="../../../../" id="documentation_options" src="../../../../_static/documentation_options.js"></script>
        <script src="../../../../_static/jquery.js"></script>
        <script src="../../../../_static/underscore.js"></script>
        <script src="../../../../_static/doctools.js"></script>
    <script src="../../../../_static/js/theme.js"></script>
    <link rel="index" title="Index" href="../../../../genindex.html" />
    <link rel="search" title="Search" href="../../../../search.html" /> 
</head>

<body class="wy-body-for-nav"> 
  <div class="wy-grid-for-nav">
    <nav data-toggle="wy-nav-shift" class="wy-nav-side">
      <div class="wy-side-scroll">
        <div class="wy-side-nav-search" >
            <a href="../../../../index.html" class="icon icon-home"> SuperGradients
          </a>
<div role="search">
  <form id="rtd-search-form" class="wy-form" action="../../../../search.html" method="get">
    <input type="text" name="q" placeholder="Search docs" />
    <input type="hidden" name="check_keywords" value="yes" />
    <input type="hidden" name="area" value="default" />
  </form>
</div>
        </div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
              <p class="caption"><span class="caption-text">SuperGradients</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html">SuperGradients</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../super_gradients.common.html">Common</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../super_gradients.training.html">Training</a></li>
</ul>

        </div>
      </div>
    </nav>

    <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
          <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
          <a href="../../../../index.html">SuperGradients</a>
      </nav>

      <div class="wy-nav-content">
        <div class="rst-content">
          <div role="navigation" aria-label="Page navigation">
  <ul class="wy-breadcrumbs">
      <li><a href="../../../../index.html" class="icon icon-home"></a> &raquo;</li>
          <li><a href="../../../index.html">Module code</a> &raquo;</li>
      <li>super_gradients.training.datasets.mixup</li>
      <li class="wy-breadcrumbs-aside">
      </li>
  </ul>
  <hr/>
</div>
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
             
  <h1>Source code for super_gradients.training.datasets.mixup</h1><div class="highlight"><pre>
<span></span><span class="sd">&quot;&quot;&quot; Mixup and Cutmix</span>

<span class="sd">Papers:</span>
<span class="sd">mixup: Beyond Empirical Risk Minimization (https://arxiv.org/abs/1710.09412)</span>

<span class="sd">CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features (https://arxiv.org/abs/1905.04899)</span>

<span class="sd">Code Reference:</span>
<span class="sd">CutMix: https://github.com/clovaai/CutMix-PyTorch</span>
<span class="sd">CutMix by timm: https://github.com/rwightman/pytorch-image-models/timm</span>

<span class="sd">&quot;&quot;&quot;</span>
<span class="kn">from</span> <span class="nn">typing</span> <span class="kn">import</span> <span class="n">List</span><span class="p">,</span> <span class="n">Union</span>

<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
<span class="kn">import</span> <span class="nn">torch</span>

<span class="kn">from</span> <span class="nn">super_gradients.training.exceptions.dataset_exceptions</span> <span class="kn">import</span> <span class="n">IllegalDatasetParameterException</span>


<div class="viewcode-block" id="one_hot"><a class="viewcode-back" href="../../../../super_gradients.training.datasets.html#super_gradients.training.datasets.mixup.one_hot">[docs]</a><span class="k">def</span> <span class="nf">one_hot</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">num_classes</span><span class="p">,</span> <span class="n">on_value</span><span class="o">=</span><span class="mf">1.</span><span class="p">,</span> <span class="n">off_value</span><span class="o">=</span><span class="mf">0.</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s1">&#39;cuda&#39;</span><span class="p">):</span>
    <span class="n">x</span> <span class="o">=</span> <span class="n">x</span><span class="o">.</span><span class="n">long</span><span class="p">()</span><span class="o">.</span><span class="n">view</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
    <span class="k">return</span> <span class="n">torch</span><span class="o">.</span><span class="n">full</span><span class="p">((</span><span class="n">x</span><span class="o">.</span><span class="n">size</span><span class="p">()[</span><span class="mi">0</span><span class="p">],</span> <span class="n">num_classes</span><span class="p">),</span> <span class="n">off_value</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span><span class="o">.</span><span class="n">scatter_</span><span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="n">x</span><span class="p">,</span> <span class="n">on_value</span><span class="p">)</span></div>


<div class="viewcode-block" id="mixup_target"><a class="viewcode-back" href="../../../../super_gradients.training.datasets.html#super_gradients.training.datasets.mixup.mixup_target">[docs]</a><span class="k">def</span> <span class="nf">mixup_target</span><span class="p">(</span><span class="n">target</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">,</span> <span class="n">num_classes</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span> <span class="n">lam</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">1.</span><span class="p">,</span> <span class="n">smoothing</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.0</span><span class="p">,</span> <span class="n">device</span><span class="p">:</span> <span class="nb">str</span> <span class="o">=</span> <span class="s1">&#39;cuda&#39;</span><span class="p">):</span>
    <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">    generate a smooth target (label) two-hot tensor to support the mixed images with different labels</span>
<span class="sd">    :param target: the targets tensor</span>
<span class="sd">    :param num_classes: number of classes (to set the final tensor size)</span>
<span class="sd">    :param lam: percentage of label a range [0, 1] in the mixing</span>
<span class="sd">    :param smoothing: the smoothing multiplier</span>
<span class="sd">    :param device: usable device [&#39;cuda&#39;, &#39;cpu&#39;]</span>
<span class="sd">    :return:</span>
<span class="sd">    &quot;&quot;&quot;</span>
    <span class="n">off_value</span> <span class="o">=</span> <span class="n">smoothing</span> <span class="o">/</span> <span class="n">num_classes</span>
    <span class="n">on_value</span> <span class="o">=</span> <span class="mf">1.</span> <span class="o">-</span> <span class="n">smoothing</span> <span class="o">+</span> <span class="n">off_value</span>
    <span class="n">y1</span> <span class="o">=</span> <span class="n">one_hot</span><span class="p">(</span><span class="n">target</span><span class="p">,</span> <span class="n">num_classes</span><span class="p">,</span> <span class="n">on_value</span><span class="o">=</span><span class="n">on_value</span><span class="p">,</span> <span class="n">off_value</span><span class="o">=</span><span class="n">off_value</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
    <span class="n">y2</span> <span class="o">=</span> <span class="n">one_hot</span><span class="p">(</span><span class="n">target</span><span class="o">.</span><span class="n">flip</span><span class="p">(</span><span class="mi">0</span><span class="p">),</span> <span class="n">num_classes</span><span class="p">,</span> <span class="n">on_value</span><span class="o">=</span><span class="n">on_value</span><span class="p">,</span> <span class="n">off_value</span><span class="o">=</span><span class="n">off_value</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="n">device</span><span class="p">)</span>
    <span class="k">return</span> <span class="n">y1</span> <span class="o">*</span> <span class="n">lam</span> <span class="o">+</span> <span class="n">y2</span> <span class="o">*</span> <span class="p">(</span><span class="mf">1.</span> <span class="o">-</span> <span class="n">lam</span><span class="p">)</span></div>


<div class="viewcode-block" id="rand_bbox"><a class="viewcode-back" href="../../../../super_gradients.training.datasets.html#super_gradients.training.datasets.mixup.rand_bbox">[docs]</a><span class="k">def</span> <span class="nf">rand_bbox</span><span class="p">(</span><span class="n">img_shape</span><span class="p">:</span> <span class="nb">tuple</span><span class="p">,</span> <span class="n">lam</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">margin</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.</span><span class="p">,</span> <span class="n">count</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="kc">None</span><span class="p">):</span>
    <span class="sd">&quot;&quot;&quot; Standard CutMix bounding-box</span>
<span class="sd">    Generates a random square bbox based on lambda value. This impl includes</span>
<span class="sd">    support for enforcing a border margin as percent of bbox dimensions.</span>

<span class="sd">    :param img_shape: Image shape as tuple</span>
<span class="sd">    :param lam: Cutmix lambda value</span>
<span class="sd">    :param margin: Percentage of bbox dimension to enforce as margin (reduce amount of box outside image)</span>
<span class="sd">    :param count: Number of bbox to generate</span>
<span class="sd">    &quot;&quot;&quot;</span>
    <span class="n">ratio</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">sqrt</span><span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">lam</span><span class="p">)</span>
    <span class="n">img_h</span><span class="p">,</span> <span class="n">img_w</span> <span class="o">=</span> <span class="n">img_shape</span><span class="p">[</span><span class="o">-</span><span class="mi">2</span><span class="p">:]</span>
    <span class="n">cut_h</span><span class="p">,</span> <span class="n">cut_w</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">img_h</span> <span class="o">*</span> <span class="n">ratio</span><span class="p">),</span> <span class="nb">int</span><span class="p">(</span><span class="n">img_w</span> <span class="o">*</span> <span class="n">ratio</span><span class="p">)</span>
    <span class="n">margin_y</span><span class="p">,</span> <span class="n">margin_x</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">margin</span> <span class="o">*</span> <span class="n">cut_h</span><span class="p">),</span> <span class="nb">int</span><span class="p">(</span><span class="n">margin</span> <span class="o">*</span> <span class="n">cut_w</span><span class="p">)</span>
    <span class="n">cy</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span> <span class="o">+</span> <span class="n">margin_y</span><span class="p">,</span> <span class="n">img_h</span> <span class="o">-</span> <span class="n">margin_y</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="n">count</span><span class="p">)</span>
    <span class="n">cx</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span> <span class="o">+</span> <span class="n">margin_x</span><span class="p">,</span> <span class="n">img_w</span> <span class="o">-</span> <span class="n">margin_x</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="n">count</span><span class="p">)</span>
    <span class="n">yl</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">cy</span> <span class="o">-</span> <span class="n">cut_h</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">img_h</span><span class="p">)</span>
    <span class="n">yh</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">cy</span> <span class="o">+</span> <span class="n">cut_h</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">img_h</span><span class="p">)</span>
    <span class="n">xl</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">cx</span> <span class="o">-</span> <span class="n">cut_w</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">img_w</span><span class="p">)</span>
    <span class="n">xh</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">clip</span><span class="p">(</span><span class="n">cx</span> <span class="o">+</span> <span class="n">cut_w</span> <span class="o">//</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">0</span><span class="p">,</span> <span class="n">img_w</span><span class="p">)</span>
    <span class="k">return</span> <span class="n">yl</span><span class="p">,</span> <span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">,</span> <span class="n">xh</span></div>


<div class="viewcode-block" id="rand_bbox_minmax"><a class="viewcode-back" href="../../../../super_gradients.training.datasets.html#super_gradients.training.datasets.mixup.rand_bbox_minmax">[docs]</a><span class="k">def</span> <span class="nf">rand_bbox_minmax</span><span class="p">(</span><span class="n">img_shape</span><span class="p">:</span> <span class="nb">tuple</span><span class="p">,</span> <span class="n">minmax</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">tuple</span><span class="p">,</span> <span class="nb">list</span><span class="p">],</span> <span class="n">count</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="kc">None</span><span class="p">):</span>
    <span class="sd">&quot;&quot;&quot; Min-Max CutMix bounding-box</span>
<span class="sd">    Inspired by Darknet cutmix impl, generates a random rectangular bbox</span>
<span class="sd">    based on min/max percent values applied to each dimension of the input image.</span>

<span class="sd">    Typical defaults for minmax are usually in the  .2-.3 for min and .8-.9 range for max.</span>

<span class="sd">    :param img_shape: Image shape as tuple</span>
<span class="sd">    :param minmax: Min and max bbox ratios (as percent of image size)</span>
<span class="sd">    :param count: Number of bbox to generate</span>
<span class="sd">    &quot;&quot;&quot;</span>
    <span class="k">assert</span> <span class="nb">len</span><span class="p">(</span><span class="n">minmax</span><span class="p">)</span> <span class="o">==</span> <span class="mi">2</span>
    <span class="n">img_h</span><span class="p">,</span> <span class="n">img_w</span> <span class="o">=</span> <span class="n">img_shape</span><span class="p">[</span><span class="o">-</span><span class="mi">2</span><span class="p">:]</span>
    <span class="n">cut_h</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">img_h</span> <span class="o">*</span> <span class="n">minmax</span><span class="p">[</span><span class="mi">0</span><span class="p">]),</span> <span class="nb">int</span><span class="p">(</span><span class="n">img_h</span> <span class="o">*</span> <span class="n">minmax</span><span class="p">[</span><span class="mi">1</span><span class="p">]),</span> <span class="n">size</span><span class="o">=</span><span class="n">count</span><span class="p">)</span>
    <span class="n">cut_w</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="nb">int</span><span class="p">(</span><span class="n">img_w</span> <span class="o">*</span> <span class="n">minmax</span><span class="p">[</span><span class="mi">0</span><span class="p">]),</span> <span class="nb">int</span><span class="p">(</span><span class="n">img_w</span> <span class="o">*</span> <span class="n">minmax</span><span class="p">[</span><span class="mi">1</span><span class="p">]),</span> <span class="n">size</span><span class="o">=</span><span class="n">count</span><span class="p">)</span>
    <span class="n">yl</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">img_h</span> <span class="o">-</span> <span class="n">cut_h</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="n">count</span><span class="p">)</span>
    <span class="n">xl</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">random</span><span class="o">.</span><span class="n">randint</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">img_w</span> <span class="o">-</span> <span class="n">cut_w</span><span class="p">,</span> <span class="n">size</span><span class="o">=</span><span class="n">count</span><span class="p">)</span>
    <span class="n">yu</span> <span class="o">=</span> <span class="n">yl</span> <span class="o">+</span> <span class="n">cut_h</span>
    <span class="n">xu</span> <span class="o">=</span> <span class="n">xl</span> <span class="o">+</span> <span class="n">cut_w</span>
    <span class="k">return</span> <span class="n">yl</span><span class="p">,</span> <span class="n">yu</span><span class="p">,</span> <span class="n">xl</span><span class="p">,</span> <span class="n">xu</span></div>


<div class="viewcode-block" id="cutmix_bbox_and_lam"><a class="viewcode-back" href="../../../../super_gradients.training.datasets.html#super_gradients.training.datasets.mixup.cutmix_bbox_and_lam">[docs]</a><span class="k">def</span> <span class="nf">cutmix_bbox_and_lam</span><span class="p">(</span><span class="n">img_shape</span><span class="p">:</span> <span class="nb">tuple</span><span class="p">,</span> <span class="n">lam</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">ratio_minmax</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">tuple</span><span class="p">,</span> <span class="nb">list</span><span class="p">]</span> <span class="o">=</span> <span class="kc">None</span><span class="p">,</span> <span class="n">correct_lam</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">True</span><span class="p">,</span>
                        <span class="n">count</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="kc">None</span><span class="p">):</span>
    <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">    Generate bbox and apply lambda correction.</span>
<span class="sd">    &quot;&quot;&quot;</span>
    <span class="k">if</span> <span class="n">ratio_minmax</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
        <span class="n">yl</span><span class="p">,</span> <span class="n">yu</span><span class="p">,</span> <span class="n">xl</span><span class="p">,</span> <span class="n">xu</span> <span class="o">=</span> <span class="n">rand_bbox_minmax</span><span class="p">(</span><span class="n">img_shape</span><span class="p">,</span> <span class="n">ratio_minmax</span><span class="p">,</span> <span class="n">count</span><span class="o">=</span><span class="n">count</span><span class="p">)</span>
    <span class="k">else</span><span class="p">:</span>
        <span class="n">yl</span><span class="p">,</span> <span class="n">yu</span><span class="p">,</span> <span class="n">xl</span><span class="p">,</span> <span class="n">xu</span> <span class="o">=</span> <span class="n">rand_bbox</span><span class="p">(</span><span class="n">img_shape</span><span class="p">,</span> <span class="n">lam</span><span class="p">,</span> <span class="n">count</span><span class="o">=</span><span class="n">count</span><span class="p">)</span>
    <span class="k">if</span> <span class="n">correct_lam</span> <span class="ow">or</span> <span class="n">ratio_minmax</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
        <span class="n">bbox_area</span> <span class="o">=</span> <span class="p">(</span><span class="n">yu</span> <span class="o">-</span> <span class="n">yl</span><span class="p">)</span> <span class="o">*</span> <span class="p">(</span><span class="n">xu</span> <span class="o">-</span> <span class="n">xl</span><span class="p">)</span>
        <span class="n">lam</span> <span class="o">=</span> <span class="mf">1.</span> <span class="o">-</span> <span class="n">bbox_area</span> <span class="o">/</span> <span class="nb">float</span><span class="p">(</span><span class="n">img_shape</span><span class="p">[</span><span class="o">-</span><span class="mi">2</span><span class="p">]</span> <span class="o">*</span> <span class="n">img_shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">])</span>
    <span class="k">return</span> <span class="p">(</span><span class="n">yl</span><span class="p">,</span> <span class="n">yu</span><span class="p">,</span> <span class="n">xl</span><span class="p">,</span> <span class="n">xu</span><span class="p">),</span> <span class="n">lam</span></div>


<div class="viewcode-block" id="CollateMixup"><a class="viewcode-back" href="../../../../super_gradients.training.datasets.html#super_gradients.training.datasets.mixup.CollateMixup">[docs]</a><span class="k">class</span> <span class="nc">CollateMixup</span><span class="p">:</span>
    <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">    Collate with Mixup/Cutmix that applies different params to each element or whole batch</span>
<span class="sd">    A Mixup impl that&#39;s performed while collating the batches.</span>
<span class="sd">    &quot;&quot;&quot;</span>

    <span class="k">def</span> <span class="fm">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">mixup_alpha</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">1.</span><span class="p">,</span> <span class="n">cutmix_alpha</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.</span><span class="p">,</span> <span class="n">cutmix_minmax</span><span class="p">:</span> <span class="n">List</span><span class="p">[</span><span class="nb">float</span><span class="p">]</span> <span class="o">=</span> <span class="kc">None</span><span class="p">,</span>
                 <span class="n">prob</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">1.0</span><span class="p">,</span> <span class="n">switch_prob</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.5</span><span class="p">,</span>
                 <span class="n">mode</span><span class="p">:</span> <span class="nb">str</span> <span class="o">=</span> <span class="s1">&#39;batch&#39;</span><span class="p">,</span> <span class="n">correct_lam</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">True</span><span class="p">,</span> <span class="n">label_smoothing</span><span class="p">:</span> <span class="nb">float</span> <span class="o">=</span> <span class="mf">0.1</span><span class="p">,</span> <span class="n">num_classes</span><span class="p">:</span> <span class="nb">int</span> <span class="o">=</span> <span class="mi">1000</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        Mixup/Cutmix that applies different params to each element or whole batch</span>

<span class="sd">        :param mixup_alpha: mixup alpha value, mixup is active if &gt; 0.</span>
<span class="sd">        :param cutmix_alpha: cutmix alpha value, cutmix is active if &gt; 0.</span>
<span class="sd">        :param cutmix_minmax: cutmix min/max image ratio, cutmix is active and uses this vs alpha if not None.</span>
<span class="sd">        :param prob: probability of applying mixup or cutmix per batch or element</span>
<span class="sd">        :param switch_prob: probability of switching to cutmix instead of mixup when both are active</span>
<span class="sd">        :param mode: how to apply mixup/cutmix params (per &#39;batch&#39;, &#39;pair&#39; (pair of elements), &#39;elem&#39; (element)</span>
<span class="sd">        :param correct_lam: apply lambda correction when cutmix bbox clipped by image borders</span>
<span class="sd">        :param label_smoothing: apply label smoothing to the mixed target tensor</span>
<span class="sd">        :param num_classes: number of classes for target</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span> <span class="o">=</span> <span class="n">mixup_alpha</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span> <span class="o">=</span> <span class="n">cutmix_alpha</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_minmax</span> <span class="o">=</span> <span class="n">cutmix_minmax</span>
        <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_minmax</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
            <span class="k">assert</span> <span class="nb">len</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">cutmix_minmax</span><span class="p">)</span> <span class="o">==</span> <span class="mi">2</span>
            <span class="c1"># force cutmix alpha == 1.0 when minmax active to keep logic simple &amp; safe</span>
            <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span> <span class="o">=</span> <span class="mf">1.0</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">mix_prob</span> <span class="o">=</span> <span class="n">prob</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">switch_prob</span> <span class="o">=</span> <span class="n">switch_prob</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">label_smoothing</span> <span class="o">=</span> <span class="n">label_smoothing</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">num_classes</span> <span class="o">=</span> <span class="n">num_classes</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">=</span> <span class="n">mode</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">correct_lam</span> <span class="o">=</span> <span class="n">correct_lam</span>  <span class="c1"># correct lambda based on clipped area for cutmix</span>
        <span class="bp">self</span><span class="o">.</span><span class="n">mixup_enabled</span> <span class="o">=</span> <span class="kc">True</span>  <span class="c1"># set to false to disable mixing (intended tp be set by train loop)</span>

    <span class="k">def</span> <span class="nf">_params_per_elem</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">batch_size</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        generate two random masks to define which elements of the batch will be mixed and how (depending on the</span>
<span class="sd">        self.mixup_enabled, self.mixup_alpha, self.cutmix_alpha parameters</span>

<span class="sd">        :param batch_size:</span>
<span class="sd">        :return: two tensors with shape=batch_size - the first contains the lambda value per batch element</span>
<span class="sd">        and the second is a binary flag indicating use of cutmix per batch element</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="n">lam</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">ones</span><span class="p">(</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
        <span class="n">use_cutmix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">zeros</span><span class="p">(</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">bool</span><span class="p">)</span>
        <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_enabled</span><span class="p">:</span>
            <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span> <span class="o">&gt;</span> <span class="mf">0.</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span> <span class="o">&gt;</span> <span class="mf">0.</span><span class="p">:</span>
                <span class="n">use_cutmix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="n">batch_size</span><span class="p">)</span> <span class="o">&lt;</span> <span class="bp">self</span><span class="o">.</span><span class="n">switch_prob</span>
                <span class="n">lam_mix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">where</span><span class="p">(</span>
                    <span class="n">use_cutmix</span><span class="p">,</span>
                    <span class="n">torch</span><span class="o">.</span><span class="n">distributions</span><span class="o">.</span><span class="n">beta</span><span class="o">.</span><span class="n">Beta</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span><span class="p">)</span><span class="o">.</span><span class="n">sample</span><span class="p">(</span><span class="n">sample_shape</span><span class="o">=</span><span class="n">batch_size</span><span class="p">),</span>
                    <span class="n">torch</span><span class="o">.</span><span class="n">distributions</span><span class="o">.</span><span class="n">beta</span><span class="o">.</span><span class="n">Beta</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span><span class="p">)</span><span class="o">.</span><span class="n">sample</span><span class="p">(</span><span class="n">sample_shape</span><span class="o">=</span><span class="n">batch_size</span><span class="p">))</span>
            <span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span> <span class="o">&gt;</span> <span class="mf">0.</span><span class="p">:</span>
                <span class="n">lam_mix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">distributions</span><span class="o">.</span><span class="n">beta</span><span class="o">.</span><span class="n">Beta</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span><span class="p">)</span><span class="o">.</span><span class="n">sample</span><span class="p">(</span><span class="n">sample_shape</span><span class="o">=</span><span class="n">batch_size</span><span class="p">)</span>
            <span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span> <span class="o">&gt;</span> <span class="mf">0.</span><span class="p">:</span>
                <span class="n">use_cutmix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">ones</span><span class="p">(</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">bool</span><span class="p">)</span>
                <span class="n">lam_mix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">distributions</span><span class="o">.</span><span class="n">beta</span><span class="o">.</span><span class="n">Beta</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span><span class="p">)</span><span class="o">.</span><span class="n">sample</span><span class="p">(</span><span class="n">sample_shape</span><span class="o">=</span><span class="n">batch_size</span><span class="p">)</span>
            <span class="k">else</span><span class="p">:</span>
                <span class="k">raise</span> <span class="n">IllegalDatasetParameterException</span><span class="p">(</span><span class="s2">&quot;One of mixup_alpha &gt; 0., cutmix_alpha &gt; 0., &quot;</span>
                                                       <span class="s2">&quot;cutmix_minmax not None should be true.&quot;</span><span class="p">)</span>
            <span class="n">lam</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">where</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="n">batch_size</span><span class="p">)</span> <span class="o">&lt;</span> <span class="bp">self</span><span class="o">.</span><span class="n">mix_prob</span><span class="p">,</span> <span class="n">lam_mix</span><span class="o">.</span><span class="n">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">),</span> <span class="n">lam</span><span class="p">)</span>
        <span class="k">return</span> <span class="n">lam</span><span class="p">,</span> <span class="n">use_cutmix</span>

    <span class="k">def</span> <span class="nf">_params_per_batch</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        generate two random parameters to define if batch will be mixed and how (depending on the</span>
<span class="sd">        self.mixup_enabled, self.mixup_alpha, self.cutmix_alpha parameters</span>

<span class="sd">        :return: two parameters - the first contains the lambda value for the whole batch</span>
<span class="sd">        and the second is a binary flag indicating use of cutmix for the batch</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="n">lam</span> <span class="o">=</span> <span class="mf">1.</span>
        <span class="n">use_cutmix</span> <span class="o">=</span> <span class="kc">False</span>

        <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_enabled</span> <span class="ow">and</span> <span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span> <span class="o">&lt;</span> <span class="bp">self</span><span class="o">.</span><span class="n">mix_prob</span><span class="p">:</span>
            <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span> <span class="o">&gt;</span> <span class="mf">0.</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span> <span class="o">&gt;</span> <span class="mf">0.</span><span class="p">:</span>
                <span class="n">use_cutmix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">rand</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span> <span class="o">&lt;</span> <span class="bp">self</span><span class="o">.</span><span class="n">switch_prob</span>
                <span class="n">lam_mix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">distributions</span><span class="o">.</span><span class="n">beta</span><span class="o">.</span><span class="n">Beta</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span><span class="p">)</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span> <span class="k">if</span> <span class="n">use_cutmix</span> <span class="k">else</span> \
                    <span class="n">torch</span><span class="o">.</span><span class="n">distributions</span><span class="o">.</span><span class="n">beta</span><span class="o">.</span><span class="n">Beta</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span><span class="p">)</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
            <span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span> <span class="o">&gt;</span> <span class="mf">0.</span><span class="p">:</span>
                <span class="n">lam_mix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">distributions</span><span class="o">.</span><span class="n">beta</span><span class="o">.</span><span class="n">Beta</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">mixup_alpha</span><span class="p">)</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
            <span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span> <span class="o">&gt;</span> <span class="mf">0.</span><span class="p">:</span>
                <span class="n">use_cutmix</span> <span class="o">=</span> <span class="kc">True</span>
                <span class="n">lam_mix</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">distributions</span><span class="o">.</span><span class="n">beta</span><span class="o">.</span><span class="n">Beta</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">cutmix_alpha</span><span class="p">)</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
            <span class="k">else</span><span class="p">:</span>
                <span class="k">raise</span> <span class="n">IllegalDatasetParameterException</span><span class="p">(</span><span class="s2">&quot;One of mixup_alpha &gt; 0., cutmix_alpha &gt; 0., &quot;</span>
                                                       <span class="s2">&quot;cutmix_minmax not None should be true.&quot;</span><span class="p">)</span>
            <span class="n">lam</span> <span class="o">=</span> <span class="nb">float</span><span class="p">(</span><span class="n">lam_mix</span><span class="p">)</span>
        <span class="k">return</span> <span class="n">lam</span><span class="p">,</span> <span class="n">use_cutmix</span>

    <span class="k">def</span> <span class="nf">_mix_elem_collate</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">output</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">,</span> <span class="n">batch</span><span class="p">:</span> <span class="nb">list</span><span class="p">,</span> <span class="n">half</span><span class="p">:</span> <span class="nb">bool</span> <span class="o">=</span> <span class="kc">False</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        This is the implementation for &#39;elem&#39; or &#39;half&#39; modes</span>
<span class="sd">        :param output: the output tensor to fill</span>
<span class="sd">        :param batch: list of thr batch items</span>
<span class="sd">        :return: a tensor containing the lambda values used for the mixing (this vector can be used for</span>
<span class="sd">        mixing the labels as well)</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="n">batch_size</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">batch</span><span class="p">)</span>
        <span class="n">num_elem</span> <span class="o">=</span> <span class="n">batch_size</span> <span class="o">//</span> <span class="mi">2</span> <span class="k">if</span> <span class="n">half</span> <span class="k">else</span> <span class="n">batch_size</span>
        <span class="k">assert</span> <span class="nb">len</span><span class="p">(</span><span class="n">output</span><span class="p">)</span> <span class="o">==</span> <span class="n">num_elem</span>
        <span class="n">lam_batch</span><span class="p">,</span> <span class="n">use_cutmix</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_params_per_elem</span><span class="p">(</span><span class="n">num_elem</span><span class="p">)</span>
        <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">num_elem</span><span class="p">):</span>
            <span class="n">j</span> <span class="o">=</span> <span class="n">batch_size</span> <span class="o">-</span> <span class="n">i</span> <span class="o">-</span> <span class="mi">1</span>
            <span class="n">lam</span> <span class="o">=</span> <span class="n">lam_batch</span><span class="p">[</span><span class="n">i</span><span class="p">]</span>
            <span class="n">mixed</span> <span class="o">=</span> <span class="n">batch</span><span class="p">[</span><span class="n">i</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span>
            <span class="k">if</span> <span class="n">lam</span> <span class="o">!=</span> <span class="mf">1.</span><span class="p">:</span>
                <span class="k">if</span> <span class="n">use_cutmix</span><span class="p">[</span><span class="n">i</span><span class="p">]:</span>
                    <span class="k">if</span> <span class="ow">not</span> <span class="n">half</span><span class="p">:</span>
                        <span class="n">mixed</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">clone</span><span class="p">(</span><span class="n">mixed</span><span class="p">)</span>
                    <span class="p">(</span><span class="n">yl</span><span class="p">,</span> <span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">,</span> <span class="n">xh</span><span class="p">),</span> <span class="n">lam</span> <span class="o">=</span> <span class="n">cutmix_bbox_and_lam</span><span class="p">(</span>
                        <span class="n">output</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="n">lam</span><span class="p">,</span> <span class="n">ratio_minmax</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">cutmix_minmax</span><span class="p">,</span> <span class="n">correct_lam</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">correct_lam</span><span class="p">)</span>
                    <span class="n">mixed</span><span class="p">[:,</span> <span class="n">yl</span><span class="p">:</span><span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">:</span><span class="n">xh</span><span class="p">]</span> <span class="o">=</span> <span class="n">batch</span><span class="p">[</span><span class="n">j</span><span class="p">][</span><span class="mi">0</span><span class="p">][:,</span> <span class="n">yl</span><span class="p">:</span><span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">:</span><span class="n">xh</span><span class="p">]</span>
                    <span class="n">lam_batch</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">=</span> <span class="n">lam</span>
                <span class="k">else</span><span class="p">:</span>
                    <span class="n">mixed</span> <span class="o">=</span> <span class="n">mixed</span> <span class="o">*</span> <span class="n">lam</span> <span class="o">+</span> <span class="n">batch</span><span class="p">[</span><span class="n">j</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">lam</span><span class="p">)</span>
            <span class="n">output</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">+=</span> <span class="n">mixed</span>
        <span class="k">if</span> <span class="n">half</span><span class="p">:</span>
            <span class="n">lam_batch</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">cat</span><span class="p">((</span><span class="n">lam_batch</span><span class="p">,</span> <span class="n">torch</span><span class="o">.</span><span class="n">ones</span><span class="p">(</span><span class="n">num_elem</span><span class="p">)))</span>
        <span class="k">return</span> <span class="n">torch</span><span class="o">.</span><span class="n">tensor</span><span class="p">(</span><span class="n">lam_batch</span><span class="p">)</span><span class="o">.</span><span class="n">unsqueeze</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>

    <span class="k">def</span> <span class="nf">_mix_pair_collate</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">output</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">,</span> <span class="n">batch</span><span class="p">:</span> <span class="nb">list</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        This is the implementation for &#39;pair&#39; mode</span>
<span class="sd">        :param output: the output tensor to fill</span>
<span class="sd">        :param batch: list of thr batch items</span>
<span class="sd">        :return: a tensor containing the lambda values used for the mixing (this vector can be used for</span>
<span class="sd">        mixing the labels as well)</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="n">batch_size</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">batch</span><span class="p">)</span>
        <span class="n">lam_batch</span><span class="p">,</span> <span class="n">use_cutmix</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_params_per_elem</span><span class="p">(</span><span class="n">batch_size</span> <span class="o">//</span> <span class="mi">2</span><span class="p">)</span>
        <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">batch_size</span> <span class="o">//</span> <span class="mi">2</span><span class="p">):</span>
            <span class="n">j</span> <span class="o">=</span> <span class="n">batch_size</span> <span class="o">-</span> <span class="n">i</span> <span class="o">-</span> <span class="mi">1</span>
            <span class="n">lam</span> <span class="o">=</span> <span class="n">lam_batch</span><span class="p">[</span><span class="n">i</span><span class="p">]</span>
            <span class="n">mixed_i</span> <span class="o">=</span> <span class="n">batch</span><span class="p">[</span><span class="n">i</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span>
            <span class="n">mixed_j</span> <span class="o">=</span> <span class="n">batch</span><span class="p">[</span><span class="n">j</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span>
            <span class="k">assert</span> <span class="mi">0</span> <span class="o">&lt;=</span> <span class="n">lam</span> <span class="o">&lt;=</span> <span class="mf">1.0</span>
            <span class="k">if</span> <span class="n">lam</span> <span class="o">&lt;</span> <span class="mf">1.</span><span class="p">:</span>
                <span class="k">if</span> <span class="n">use_cutmix</span><span class="p">[</span><span class="n">i</span><span class="p">]:</span>
                    <span class="p">(</span><span class="n">yl</span><span class="p">,</span> <span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">,</span> <span class="n">xh</span><span class="p">),</span> <span class="n">lam</span> <span class="o">=</span> <span class="n">cutmix_bbox_and_lam</span><span class="p">(</span>
                        <span class="n">output</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="n">lam</span><span class="p">,</span> <span class="n">ratio_minmax</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">cutmix_minmax</span><span class="p">,</span> <span class="n">correct_lam</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">correct_lam</span><span class="p">)</span>
                    <span class="n">patch_i</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">clone</span><span class="p">(</span><span class="n">mixed_i</span><span class="p">[:,</span> <span class="n">yl</span><span class="p">:</span><span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">:</span><span class="n">xh</span><span class="p">])</span>
                    <span class="n">mixed_i</span><span class="p">[:,</span> <span class="n">yl</span><span class="p">:</span><span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">:</span><span class="n">xh</span><span class="p">]</span> <span class="o">=</span> <span class="n">mixed_j</span><span class="p">[:,</span> <span class="n">yl</span><span class="p">:</span><span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">:</span><span class="n">xh</span><span class="p">]</span>
                    <span class="n">mixed_j</span><span class="p">[:,</span> <span class="n">yl</span><span class="p">:</span><span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">:</span><span class="n">xh</span><span class="p">]</span> <span class="o">=</span> <span class="n">patch_i</span>
                    <span class="n">lam_batch</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">=</span> <span class="n">lam</span>
                <span class="k">else</span><span class="p">:</span>
                    <span class="n">mixed_temp</span> <span class="o">=</span> <span class="n">mixed_i</span><span class="o">.</span><span class="n">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span> <span class="o">*</span> <span class="n">lam</span> <span class="o">+</span> <span class="n">mixed_j</span><span class="o">.</span><span class="n">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span> <span class="o">*</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">lam</span><span class="p">)</span>
                    <span class="n">mixed_j</span> <span class="o">=</span> <span class="n">mixed_j</span><span class="o">.</span><span class="n">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span> <span class="o">*</span> <span class="n">lam</span> <span class="o">+</span> <span class="n">mixed_i</span><span class="o">.</span><span class="n">type</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span> <span class="o">*</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">lam</span><span class="p">)</span>
                    <span class="n">mixed_i</span> <span class="o">=</span> <span class="n">mixed_temp</span>
                    <span class="n">torch</span><span class="o">.</span><span class="n">rint</span><span class="p">(</span><span class="n">mixed_j</span><span class="p">,</span> <span class="n">out</span><span class="o">=</span><span class="n">mixed_j</span><span class="p">)</span>
                    <span class="n">torch</span><span class="o">.</span><span class="n">rint</span><span class="p">(</span><span class="n">mixed_i</span><span class="p">,</span> <span class="n">out</span><span class="o">=</span><span class="n">mixed_i</span><span class="p">)</span>
            <span class="n">output</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">+=</span> <span class="n">mixed_i</span>
            <span class="n">output</span><span class="p">[</span><span class="n">j</span><span class="p">]</span> <span class="o">+=</span> <span class="n">mixed_j</span>
        <span class="n">lam_batch</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">cat</span><span class="p">((</span><span class="n">lam_batch</span><span class="p">,</span> <span class="n">lam_batch</span><span class="p">[::</span><span class="o">-</span><span class="mi">1</span><span class="p">]))</span>
        <span class="k">return</span> <span class="n">torch</span><span class="o">.</span><span class="n">tensor</span><span class="p">(</span><span class="n">lam_batch</span><span class="p">)</span><span class="o">.</span><span class="n">unsqueeze</span><span class="p">(</span><span class="mi">1</span><span class="p">)</span>

    <span class="k">def</span> <span class="nf">_mix_batch_collate</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">output</span><span class="p">:</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">,</span> <span class="n">batch</span><span class="p">:</span> <span class="nb">list</span><span class="p">):</span>
        <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">        This is the implementation for &#39;batch&#39; mode</span>
<span class="sd">        :param output: the output tensor to fill</span>
<span class="sd">        :param batch: list of thr batch items</span>
<span class="sd">        :return: the lambda value used for the mixing</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="n">batch_size</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">batch</span><span class="p">)</span>
        <span class="n">lam</span><span class="p">,</span> <span class="n">use_cutmix</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_params_per_batch</span><span class="p">()</span>
        <span class="k">if</span> <span class="n">use_cutmix</span><span class="p">:</span>
            <span class="p">(</span><span class="n">yl</span><span class="p">,</span> <span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">,</span> <span class="n">xh</span><span class="p">),</span> <span class="n">lam</span> <span class="o">=</span> <span class="n">cutmix_bbox_and_lam</span><span class="p">(</span>
                <span class="n">output</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="n">lam</span><span class="p">,</span> <span class="n">ratio_minmax</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">cutmix_minmax</span><span class="p">,</span> <span class="n">correct_lam</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">correct_lam</span><span class="p">)</span>
        <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">batch_size</span><span class="p">):</span>
            <span class="n">j</span> <span class="o">=</span> <span class="n">batch_size</span> <span class="o">-</span> <span class="n">i</span> <span class="o">-</span> <span class="mi">1</span>
            <span class="n">mixed</span> <span class="o">=</span> <span class="n">batch</span><span class="p">[</span><span class="n">i</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span>
            <span class="k">if</span> <span class="n">lam</span> <span class="o">!=</span> <span class="mf">1.</span><span class="p">:</span>
                <span class="k">if</span> <span class="n">use_cutmix</span><span class="p">:</span>
                    <span class="n">mixed</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">clone</span><span class="p">(</span><span class="n">mixed</span><span class="p">)</span>  <span class="c1"># don&#39;t want to modify the original while iterating</span>
                    <span class="n">mixed</span><span class="p">[:,</span> <span class="n">yl</span><span class="p">:</span><span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">:</span><span class="n">xh</span><span class="p">]</span> <span class="o">=</span> <span class="n">batch</span><span class="p">[</span><span class="n">j</span><span class="p">][</span><span class="mi">0</span><span class="p">][:,</span> <span class="n">yl</span><span class="p">:</span><span class="n">yh</span><span class="p">,</span> <span class="n">xl</span><span class="p">:</span><span class="n">xh</span><span class="p">]</span>
                <span class="k">else</span><span class="p">:</span>
                    <span class="n">mixed</span> <span class="o">=</span> <span class="n">mixed</span> <span class="o">*</span> <span class="n">lam</span> <span class="o">+</span> <span class="n">batch</span><span class="p">[</span><span class="n">j</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span> <span class="o">*</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">lam</span><span class="p">)</span>
            <span class="n">output</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="o">+=</span> <span class="n">mixed</span>
        <span class="k">return</span> <span class="n">lam</span>

    <span class="k">def</span> <span class="fm">__call__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">batch</span><span class="p">,</span> <span class="n">_</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
        <span class="n">batch_size</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">batch</span><span class="p">)</span>
        <span class="k">if</span> <span class="n">batch_size</span> <span class="o">%</span> <span class="mi">2</span> <span class="o">!=</span> <span class="mi">0</span><span class="p">:</span>
            <span class="k">raise</span> <span class="n">IllegalDatasetParameterException</span><span class="p">(</span><span class="s1">&#39;Batch size should be even when using this&#39;</span><span class="p">)</span>
        <span class="n">half</span> <span class="o">=</span> <span class="s1">&#39;half&#39;</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">mode</span>
        <span class="k">if</span> <span class="n">half</span><span class="p">:</span>
            <span class="n">batch_size</span> <span class="o">//=</span> <span class="mi">2</span>
        <span class="n">output</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="n">batch_size</span><span class="p">,</span> <span class="o">*</span><span class="n">batch</span><span class="p">[</span><span class="mi">0</span><span class="p">][</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">shape</span><span class="p">),</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">float32</span><span class="p">)</span>
        <span class="k">if</span> <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">==</span> <span class="s1">&#39;elem&#39;</span> <span class="ow">or</span> <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">==</span> <span class="s1">&#39;half&#39;</span><span class="p">:</span>
            <span class="n">lam</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mix_elem_collate</span><span class="p">(</span><span class="n">output</span><span class="p">,</span> <span class="n">batch</span><span class="p">,</span> <span class="n">half</span><span class="o">=</span><span class="n">half</span><span class="p">)</span>
        <span class="k">elif</span> <span class="bp">self</span><span class="o">.</span><span class="n">mode</span> <span class="o">==</span> <span class="s1">&#39;pair&#39;</span><span class="p">:</span>
            <span class="n">lam</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mix_pair_collate</span><span class="p">(</span><span class="n">output</span><span class="p">,</span> <span class="n">batch</span><span class="p">)</span>
        <span class="k">else</span><span class="p">:</span>
            <span class="n">lam</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">_mix_batch_collate</span><span class="p">(</span><span class="n">output</span><span class="p">,</span> <span class="n">batch</span><span class="p">)</span>
        <span class="n">target</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">tensor</span><span class="p">([</span><span class="n">b</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span> <span class="k">for</span> <span class="n">b</span> <span class="ow">in</span> <span class="n">batch</span><span class="p">],</span> <span class="n">dtype</span><span class="o">=</span><span class="n">torch</span><span class="o">.</span><span class="n">int32</span><span class="p">)</span>
        <span class="n">target</span> <span class="o">=</span> <span class="n">mixup_target</span><span class="p">(</span><span class="n">target</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">num_classes</span><span class="p">,</span> <span class="n">lam</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">label_smoothing</span><span class="p">,</span> <span class="n">device</span><span class="o">=</span><span class="s1">&#39;cpu&#39;</span><span class="p">)</span>
        <span class="n">target</span> <span class="o">=</span> <span class="n">target</span><span class="p">[:</span><span class="n">batch_size</span><span class="p">]</span>

        <span class="k">return</span> <span class="n">output</span><span class="p">,</span> <span class="n">target</span></div>
</pre></div>

           </div>
          </div>
          <footer>

  <hr/>

  <div role="contentinfo">
    <p>&#169; Copyright 2021, SuperGradients team.</p>
  </div>

  Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
    <a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
    provided by <a href="https://readthedocs.org">Read the Docs</a>.
   

</footer>
        </div>
      </div>
    </section>
  </div>
  <script>
      jQuery(function () {
          SphinxRtdTheme.Navigation.enable(true);
      });
  </script> 

</body>
</html>