#396 Trainer constructor cleanup

Merged
Ghost merged 1 commits into Deci-AI:master from deci-ai:feature/SG-266_clean_trainer_ctor
  
    
        
          
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

          
            <!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.utils.segmentation_utils &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">Welcome To SuperGradients</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html">Fill our 4-question quick survey! We will raffle free SuperGradients swag between those who will participate -&gt; Fill Survey</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../welcome.html#supergradients">SuperGradients</a></li>
</ul>
<p class="caption"><span class="caption-text">Technical Documentation</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../../super_gradients.common.html">Common package</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../super_gradients.training.html">Training package</a></li>
</ul>
<p class="caption"><span class="caption-text">User Guide</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html">What is SuperGradients?</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#introducing-the-supergradients-library">Introducing the SuperGradients library</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#installation">Installation</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#integrating-your-training-code-complete-walkthrough">Integrating Your Training Code - Complete Walkthrough</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#training-parameters">Training Parameters</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#logs-and-checkpoints">Logs and Checkpoints</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#dataset-parameters">Dataset Parameters</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#network-architectures">Network Architectures</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#pretrained-models">Pretrained Models</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#how-to-reproduce-our-training-recipes">How To Reproduce Our Training Recipes</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#professional-tools-integration">Professional Tools Integration</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../../user_guide.html#supergradients-faq">SuperGradients FAQ</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.utils.segmentation_utils</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.utils.segmentation_utils</h1><div class="highlight"><pre>
<span></span><span class="kn">import</span> <span class="nn">os</span>
<span class="kn">import</span> <span class="nn">cv2</span>
<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
<span class="kn">from</span> <span class="nn">typing</span> <span class="kn">import</span> <span class="n">Union</span><span class="p">,</span> <span class="n">Callable</span>
<span class="kn">import</span> <span class="nn">torch</span>
<span class="kn">import</span> <span class="nn">torch.nn.functional</span> <span class="k">as</span> <span class="nn">F</span>
<span class="kn">from</span> <span class="nn">torchvision.utils</span> <span class="kn">import</span> <span class="n">draw_segmentation_masks</span>
<span class="c1"># FIXME: REFACTOR AUGMENTATIONS, CONSIDER USING A MORE EFFICIENT LIBRARIES SUCH AS, IMGAUG, DALI ETC.</span>
<span class="kn">from</span> <span class="nn">super_gradients.training</span> <span class="kn">import</span> <span class="n">utils</span> <span class="k">as</span> <span class="n">core_utils</span>
<div class="viewcode-block" id="coco_sub_classes_inclusion_tuples_list"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.segmentation_utils.coco_sub_classes_inclusion_tuples_list">[docs]</a><span class="k">def</span> <span class="nf">coco_sub_classes_inclusion_tuples_list</span><span class="p">():</span>
    <span class="k">return</span> <span class="p">[(</span><span class="mi">0</span><span class="p">,</span> <span class="s1">&#39;background&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">5</span><span class="p">,</span> <span class="s1">&#39;airplane&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">2</span><span class="p">,</span> <span class="s1">&#39;bicycle&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">16</span><span class="p">,</span> <span class="s1">&#39;bird&#39;</span><span class="p">),</span>
            <span class="p">(</span><span class="mi">9</span><span class="p">,</span> <span class="s1">&#39;boat&#39;</span><span class="p">),</span>
            <span class="p">(</span><span class="mi">44</span><span class="p">,</span> <span class="s1">&#39;bottle&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">6</span><span class="p">,</span> <span class="s1">&#39;bus&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">3</span><span class="p">,</span> <span class="s1">&#39;car&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">17</span><span class="p">,</span> <span class="s1">&#39;cat&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">62</span><span class="p">,</span> <span class="s1">&#39;chair&#39;</span><span class="p">),</span>
            <span class="p">(</span><span class="mi">21</span><span class="p">,</span> <span class="s1">&#39;cow&#39;</span><span class="p">),</span>
            <span class="p">(</span><span class="mi">67</span><span class="p">,</span> <span class="s1">&#39;dining table&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">18</span><span class="p">,</span> <span class="s1">&#39;dog&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">19</span><span class="p">,</span> <span class="s1">&#39;horse&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">4</span><span class="p">,</span> <span class="s1">&#39;motorcycle&#39;</span><span class="p">),</span>
            <span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="s1">&#39;person&#39;</span><span class="p">),</span>
            <span class="p">(</span><span class="mi">64</span><span class="p">,</span> <span class="s1">&#39;potted plant&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">20</span><span class="p">,</span> <span class="s1">&#39;sheep&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">63</span><span class="p">,</span> <span class="s1">&#39;couch&#39;</span><span class="p">),</span> <span class="p">(</span><span class="mi">7</span><span class="p">,</span> <span class="s1">&#39;train&#39;</span><span class="p">),</span>
            <span class="p">(</span><span class="mi">72</span><span class="p">,</span> <span class="s1">&#39;tv&#39;</span><span class="p">)]</span></div>
<div class="viewcode-block" id="to_one_hot"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.segmentation_utils.to_one_hot">[docs]</a><span class="k">def</span> <span class="nf">to_one_hot</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">ignore_index</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">    Target label to one_hot tensor. labels and ignore_index must be consecutive numbers.</span>
<span class="sd">    :param target: Class labels long tensor, with shape [N, H, W]</span>
<span class="sd">    :param num_classes: num of classes in datasets excluding ignore label, this is the output channels of the one hot</span>
<span class="sd">        result.</span>
<span class="sd">    :return: one hot tensor with shape [N, num_classes, H, W]</span>
<span class="sd">    &quot;&quot;&quot;</span>
    <span class="n">num_classes</span> <span class="o">=</span> <span class="n">num_classes</span> <span class="k">if</span> <span class="n">ignore_index</span> <span class="ow">is</span> <span class="kc">None</span> <span class="k">else</span> <span class="n">num_classes</span> <span class="o">+</span> <span class="mi">1</span>
    <span class="n">one_hot</span> <span class="o">=</span> <span class="n">F</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="o">.</span><span class="n">permute</span><span class="p">((</span><span class="mi">0</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">1</span><span class="p">,</span> <span class="mi">2</span><span class="p">))</span>
    <span class="k">if</span> <span class="n">ignore_index</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
        <span class="c1"># remove ignore_index channel</span>
        <span class="n">one_hot</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">one_hot</span><span class="p">[:,</span> <span class="p">:</span><span class="n">ignore_index</span><span class="p">],</span> <span class="n">one_hot</span><span class="p">[:,</span> <span class="n">ignore_index</span> <span class="o">+</span> <span class="mi">1</span><span class="p">:]],</span> <span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
    <span class="k">return</span> <span class="n">one_hot</span></div>
<div class="viewcode-block" id="reverse_imagenet_preprocessing"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.segmentation_utils.reverse_imagenet_preprocessing">[docs]</a><span class="k">def</span> <span class="nf">reverse_imagenet_preprocessing</span><span class="p">(</span><span class="n">im_tensor</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="o">-&gt;</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">:</span>
    <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">    :param im_tensor: images in a batch after preprocessing for inference, RGB, (B, C, H, W)</span>
<span class="sd">    :return:          images in a batch in cv2 format, BGR, (B, H, W, C)</span>
<span class="sd">    &quot;&quot;&quot;</span>
    <span class="n">im_np</span> <span class="o">=</span> <span class="n">im_tensor</span><span class="o">.</span><span class="n">cpu</span><span class="p">()</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
    <span class="n">im_np</span> <span class="o">=</span> <span class="n">im_np</span><span class="p">[:,</span> <span class="p">::</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="p">:,</span> <span class="p">:]</span><span class="o">.</span><span class="n">transpose</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
    <span class="n">im_np</span> <span class="o">*=</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">([[[</span><span class="mf">.229</span><span class="p">,</span> <span class="mf">.224</span><span class="p">,</span> <span class="mf">.225</span><span class="p">][::</span><span class="o">-</span><span class="mi">1</span><span class="p">]]])</span>
    <span class="n">im_np</span> <span class="o">+=</span> <span class="n">np</span><span class="o">.</span><span class="n">array</span><span class="p">([[[</span><span class="mf">.485</span><span class="p">,</span> <span class="mf">.456</span><span class="p">,</span> <span class="mf">.406</span><span class="p">][::</span><span class="o">-</span><span class="mi">1</span><span class="p">]]])</span>
    <span class="n">im_np</span> <span class="o">*=</span> <span class="mf">255.</span>
    <span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">ascontiguousarray</span><span class="p">(</span><span class="n">im_np</span><span class="p">,</span> <span class="n">dtype</span><span class="o">=</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">)</span></div>
<div class="viewcode-block" id="BinarySegmentationVisualization"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.segmentation_utils.BinarySegmentationVisualization">[docs]</a><span class="k">class</span> <span class="nc">BinarySegmentationVisualization</span><span class="p">:</span>
    <span class="nd">@staticmethod</span>
    <span class="k">def</span> <span class="nf">_visualize_image</span><span class="p">(</span><span class="n">image_np</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">,</span> <span class="n">pred_mask</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">target_mask</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">image_scale</span><span class="p">:</span> <span class="nb">float</span><span class="p">,</span> <span class="n">checkpoint_dir</span><span class="p">:</span> <span class="nb">str</span><span class="p">,</span> <span class="n">image_name</span><span class="p">:</span> <span class="nb">str</span><span class="p">):</span>
        <span class="n">pred_mask</span> <span class="o">=</span> <span class="n">pred_mask</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span>
        <span class="n">image_np</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">from_numpy</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">moveaxis</span><span class="p">(</span><span class="n">image_np</span><span class="p">,</span> <span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="mi">0</span><span class="p">)</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">))</span>
        <span class="n">pred_mask</span> <span class="o">=</span> <span class="n">pred_mask</span><span class="p">[</span><span class="n">np</span><span class="o">.</span><span class="n">newaxis</span><span class="p">,</span> <span class="p">:,</span> <span class="p">:]</span> <span class="o">&gt;</span> <span class="mf">0.5</span>
        <span class="n">target_mask</span> <span class="o">=</span> <span class="n">target_mask</span><span class="p">[</span><span class="n">np</span><span class="o">.</span><span class="n">newaxis</span><span class="p">,</span> <span class="p">:,</span> <span class="p">:]</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="nb">bool</span><span class="p">)</span>
        <span class="n">tp_mask</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">logical_and</span><span class="p">(</span><span class="n">pred_mask</span><span class="p">,</span> <span class="n">target_mask</span><span class="p">)</span>
        <span class="n">fp_mask</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">logical_and</span><span class="p">(</span><span class="n">pred_mask</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">logical_not</span><span class="p">(</span><span class="n">target_mask</span><span class="p">))</span>
        <span class="n">fn_mask</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">logical_and</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">logical_not</span><span class="p">(</span><span class="n">pred_mask</span><span class="p">),</span> <span class="n">target_mask</span><span class="p">)</span>
        <span class="n">overlay</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">from_numpy</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">concatenate</span><span class="p">([</span><span class="n">tp_mask</span><span class="p">,</span> <span class="n">fp_mask</span><span class="p">,</span> <span class="n">fn_mask</span><span class="p">]))</span>
        <span class="c1"># SWITCH BETWEEN BLUE AND RED IF WE SAVE THE IMAGE ON THE DISC AS OTHERWISE WE CHANGE CHANNEL ORDERING</span>
        <span class="n">colors</span> <span class="o">=</span> <span class="p">[</span><span class="s1">&#39;green&#39;</span><span class="p">,</span> <span class="s1">&#39;red&#39;</span><span class="p">,</span> <span class="s1">&#39;blue&#39;</span><span class="p">]</span>
        <span class="n">res_image</span> <span class="o">=</span> <span class="n">draw_segmentation_masks</span><span class="p">(</span><span class="n">image_np</span><span class="p">,</span> <span class="n">overlay</span><span class="p">,</span> <span class="n">colors</span><span class="o">=</span><span class="n">colors</span><span class="p">)</span><span class="o">.</span><span class="n">detach</span><span class="p">()</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
        <span class="n">res_image</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">concatenate</span><span class="p">([</span><span class="n">res_image</span><span class="p">[</span><span class="n">ch</span><span class="p">,</span> <span class="p">:,</span> <span class="p">:,</span> <span class="n">np</span><span class="o">.</span><span class="n">newaxis</span><span class="p">]</span> <span class="k">for</span> <span class="n">ch</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">3</span><span class="p">)],</span> <span class="mi">2</span><span class="p">)</span>
        <span class="n">res_image</span> <span class="o">=</span> <span class="n">cv2</span><span class="o">.</span><span class="n">resize</span><span class="p">(</span><span class="n">res_image</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">uint8</span><span class="p">),</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="n">fx</span><span class="o">=</span><span class="n">image_scale</span><span class="p">,</span> <span class="n">fy</span><span class="o">=</span><span class="n">image_scale</span><span class="p">,</span>
                               <span class="n">interpolation</span><span class="o">=</span><span class="n">cv2</span><span class="o">.</span><span class="n">INTER_NEAREST</span><span class="p">)</span>
        <span class="k">if</span> <span class="n">checkpoint_dir</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
            <span class="k">return</span> <span class="n">res_image</span>
        <span class="k">else</span><span class="p">:</span>
            <span class="n">cv2</span><span class="o">.</span><span class="n">imwrite</span><span class="p">(</span><span class="n">os</span><span class="o">.</span><span class="n">path</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">checkpoint_dir</span><span class="p">,</span> <span class="nb">str</span><span class="p">(</span><span class="n">image_name</span><span class="p">)</span> <span class="o">+</span> <span class="s1">&#39;.jpg&#39;</span><span class="p">),</span> <span class="n">res_image</span><span class="p">)</span>
<div class="viewcode-block" id="BinarySegmentationVisualization.visualize_batch"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.segmentation_utils.BinarySegmentationVisualization.visualize_batch">[docs]</a>    <span class="nd">@staticmethod</span>
    <span class="k">def</span> <span class="nf">visualize_batch</span><span class="p">(</span><span class="n">image_tensor</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">pred_mask</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">target_mask</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_name</span><span class="p">:</span> <span class="n">Union</span><span class="p">[</span><span class="nb">int</span><span class="p">,</span> <span class="nb">str</span><span class="p">],</span> <span class="n">checkpoint_dir</span><span class="p">:</span> <span class="nb">str</span> <span class="o">=</span> <span class="kc">None</span><span class="p">,</span>
                        <span class="n">undo_preprocessing_func</span><span class="p">:</span> <span class="n">Callable</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">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">]</span> <span class="o">=</span> <span class="n">reverse_imagenet_preprocessing</span><span class="p">,</span>
                        <span class="n">image_scale</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="sd">&quot;&quot;&quot;</span>
<span class="sd">        A helper function to visualize detections predicted by a network:</span>
<span class="sd">        saves images into a given path with a name that is {batch_name}_{imade_idx_in_the_batch}.jpg, one batch per call.</span>
<span class="sd">        Colors are generated on the fly: uniformly sampled from color wheel to support all given classes.</span>
<span class="sd">        :param image_tensor:            rgb images, (B, H, W, 3)</span>
<span class="sd">        :param pred_boxes:              boxes after NMS for each image in a batch, each (Num_boxes, 6),</span>
<span class="sd">                                        values on dim 1 are: x1, y1, x2, y2, confidence, class</span>
<span class="sd">        :param target_boxes:            (Num_targets, 6), values on dim 1 are: image id in a batch, class, x y w h</span>
<span class="sd">                                        (coordinates scaled to [0, 1])</span>
<span class="sd">        :param batch_name:              id of the current batch to use for image naming</span>
<span class="sd">        :param checkpoint_dir:          a path where images with boxes will be saved. if None, the result images will</span>
<span class="sd">                                        be returns as a list of numpy image arrays</span>
<span class="sd">        :param undo_preprocessing_func: a function to convert preprocessed images tensor into a batch of cv2-like images</span>
<span class="sd">        :param image_scale:             scale factor for output image</span>
<span class="sd">        &quot;&quot;&quot;</span>
        <span class="n">image_np</span> <span class="o">=</span> <span class="n">undo_preprocessing_func</span><span class="p">(</span><span class="n">image_tensor</span><span class="o">.</span><span class="n">detach</span><span class="p">())</span>
        <span class="n">pred_mask</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">sigmoid</span><span class="p">(</span><span class="n">pred_mask</span><span class="p">[:,</span> <span class="mi">0</span><span class="p">,</span> <span class="p">:,</span> <span class="p">:])</span>  <span class="c1"># comment out</span>
        <span class="n">out_images</span> <span class="o">=</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">image_np</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]):</span>
            <span class="n">preds</span> <span class="o">=</span> <span class="n">pred_mask</span><span class="p">[</span><span class="n">i</span><span class="p">]</span><span class="o">.</span><span class="n">detach</span><span class="p">()</span><span class="o">.</span><span class="n">cpu</span><span class="p">()</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
            <span class="n">targets</span> <span class="o">=</span> <span class="n">target_mask</span><span class="p">[</span><span class="n">i</span><span class="p">]</span><span class="o">.</span><span class="n">detach</span><span class="p">()</span><span class="o">.</span><span class="n">cpu</span><span class="p">()</span><span class="o">.</span><span class="n">numpy</span><span class="p">()</span>
            <span class="n">image_name</span> <span class="o">=</span> <span class="s1">&#39;_&#39;</span><span class="o">.</span><span class="n">join</span><span class="p">([</span><span class="nb">str</span><span class="p">(</span><span class="n">batch_name</span><span class="p">),</span> <span class="nb">str</span><span class="p">(</span><span class="n">i</span><span class="p">)])</span>
            <span class="n">res_image</span> <span class="o">=</span> <span class="n">BinarySegmentationVisualization</span><span class="o">.</span><span class="n">_visualize_image</span><span class="p">(</span><span class="n">image_np</span><span class="p">[</span><span class="n">i</span><span class="p">],</span> <span class="n">preds</span><span class="p">,</span> <span class="n">targets</span><span class="p">,</span> <span class="n">image_scale</span><span class="p">,</span>
                                                                         <span class="n">checkpoint_dir</span><span class="p">,</span> <span class="n">image_name</span><span class="p">)</span>
            <span class="k">if</span> <span class="n">res_image</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
                <span class="n">out_images</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">res_image</span><span class="p">)</span>
        <span class="k">return</span> <span class="n">out_images</span></div></div>
<div class="viewcode-block" id="visualize_batches"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.segmentation_utils.visualize_batches">[docs]</a><span class="k">def</span> <span class="nf">visualize_batches</span><span class="p">(</span><span class="n">dataloader</span><span class="p">,</span> <span class="n">module</span><span class="p">,</span> <span class="n">visualization_path</span><span class="p">,</span> <span class="n">num_batches</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">undo_preprocessing_func</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
    <span class="n">os</span><span class="o">.</span><span class="n">makedirs</span><span class="p">(</span><span class="n">visualization_path</span><span class="p">,</span> <span class="n">exist_ok</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
    <span class="k">for</span> <span class="n">batch_i</span><span class="p">,</span> <span class="p">(</span><span class="n">imgs</span><span class="p">,</span> <span class="n">targets</span><span class="p">)</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">dataloader</span><span class="p">):</span>
        <span class="k">if</span> <span class="n">batch_i</span> <span class="o">==</span> <span class="n">num_batches</span><span class="p">:</span>
            <span class="k">return</span>
        <span class="n">imgs</span> <span class="o">=</span> <span class="n">core_utils</span><span class="o">.</span><span class="n">tensor_container_to_device</span><span class="p">(</span><span class="n">imgs</span><span class="p">,</span> <span class="n">torch</span><span class="o">.</span><span class="n">device</span><span class="p">(</span><span class="s1">&#39;cuda:0&#39;</span><span class="p">))</span>
        <span class="n">targets</span> <span class="o">=</span> <span class="n">core_utils</span><span class="o">.</span><span class="n">tensor_container_to_device</span><span class="p">(</span><span class="n">targets</span><span class="p">,</span> <span class="n">torch</span><span class="o">.</span><span class="n">device</span><span class="p">(</span><span class="s1">&#39;cuda:0&#39;</span><span class="p">))</span>
        <span class="n">pred_mask</span> <span class="o">=</span> <span class="n">module</span><span class="p">(</span><span class="n">imgs</span><span class="p">)</span>
        <span class="c1"># Visualize the batch</span>
        <span class="k">if</span> <span class="n">undo_preprocessing_func</span><span class="p">:</span>
            <span class="n">BinarySegmentationVisualization</span><span class="o">.</span><span class="n">visualize_batch</span><span class="p">(</span><span class="n">imgs</span><span class="p">,</span> <span class="n">pred_mask</span><span class="p">,</span> <span class="n">targets</span><span class="p">,</span> <span class="n">batch_i</span><span class="p">,</span> <span class="n">visualization_path</span><span class="p">,</span>
                                                            <span class="n">undo_preprocessing_func</span><span class="o">=</span><span class="n">undo_preprocessing_func</span><span class="p">)</span>
        <span class="k">else</span><span class="p">:</span>
            <span class="n">BinarySegmentationVisualization</span><span class="o">.</span><span class="n">visualize_batch</span><span class="p">(</span><span class="n">imgs</span><span class="p">,</span> <span class="n">pred_mask</span><span class="p">,</span> <span class="n">targets</span><span class="p">,</span> <span class="n">batch_i</span><span class="p">,</span> <span class="n">visualization_path</span><span class="p">)</span></div>
<div class="viewcode-block" id="one_hot_to_binary_edge"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.segmentation_utils.one_hot_to_binary_edge">[docs]</a><span class="k">def</span> <span class="nf">one_hot_to_binary_edge</span><span class="p">(</span><span class="n">x</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">kernel_size</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span>
                           <span class="n">flatten_channels</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="o">-&gt;</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">:</span>
    <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">    Utils function to create edge feature maps.</span>
<span class="sd">    :param x: input tensor, must be one_hot tensor with shape [B, C, H, W]</span>
<span class="sd">    :param kernel_size: kernel size of dilation erosion convolutions. The result edge widths depends on this argument as</span>
<span class="sd">        follows: `edge_width = kernel - 1`</span>
<span class="sd">    :param flatten_channels: Whether to apply logical_or across channels dimension, if at least one pixel class is</span>
<span class="sd">        considered as edge pixel flatten value is 1. If set as `False` the output tensor shape is [B, C, H, W], else</span>
<span class="sd">        [B, 1, H, W]. Default is `True`.</span>
<span class="sd">    :return: one_hot edge torch.Tensor.</span>
<span class="sd">    &quot;&quot;&quot;</span>
    <span class="k">if</span> <span class="n">kernel_size</span> <span class="o">&lt;</span> <span class="mi">0</span> <span class="ow">or</span> <span class="n">kernel_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="ne">ValueError</span><span class="p">(</span><span class="sa">f</span><span class="s2">&quot;kernel size must be an odd positive values, such as [1, 3, 5, ..], found: </span><span class="si">{</span><span class="n">kernel_size</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">)</span>
    <span class="n">_kernel</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">x</span><span class="o">.</span><span class="n">size</span><span class="p">(</span><span class="mi">1</span><span class="p">),</span> <span class="mi">1</span><span class="p">,</span> <span class="n">kernel_size</span><span class="p">,</span> <span class="n">kernel_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">device</span><span class="o">=</span><span class="n">x</span><span class="o">.</span><span class="n">device</span><span class="p">)</span>
    <span class="n">padding</span> <span class="o">=</span> <span class="p">(</span><span class="n">kernel_size</span> <span class="o">-</span> <span class="mi">1</span><span class="p">)</span> <span class="o">//</span> <span class="mi">2</span>
    <span class="c1"># Use replicate padding to prevent class shifting and edge formation at the image boundaries.</span>
    <span class="n">padded_x</span> <span class="o">=</span> <span class="n">F</span><span class="o">.</span><span class="n">pad</span><span class="p">(</span><span class="n">x</span><span class="o">.</span><span class="n">float</span><span class="p">(),</span> <span class="n">mode</span><span class="o">=</span><span class="s2">&quot;replicate&quot;</span><span class="p">,</span> <span class="n">pad</span><span class="o">=</span><span class="p">[</span><span class="n">padding</span><span class="p">]</span> <span class="o">*</span> <span class="mi">4</span><span class="p">)</span>
    <span class="c1"># The binary edges feature map is created by subtracting dilated features from erosed features.</span>
    <span class="c1"># First the positive one value masks are expanded (dilation) by applying a sliding window filter of one values.</span>
    <span class="c1"># The resulted output is then clamped to binary format to [0, 1], this way the one-hot boundaries are expanded by</span>
    <span class="c1"># (kernel_size - 1) / 2.</span>
    <span class="n">dilation</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">clamp</span><span class="p">(</span>
        <span class="n">F</span><span class="o">.</span><span class="n">conv2d</span><span class="p">(</span><span class="n">padded_x</span><span class="p">,</span> <span class="n">_kernel</span><span class="p">,</span> <span class="n">groups</span><span class="o">=</span><span class="n">x</span><span class="o">.</span><span class="n">size</span><span class="p">(</span><span class="mi">1</span><span class="p">)),</span>
        <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span>
    <span class="p">)</span>
    <span class="c1"># Similar to dilation, erosion (can be seen as inverse of dilation) is applied to contract the one-hot features by</span>
    <span class="c1"># applying a dilation operation on the inverse of the one-hot features.</span>
    <span class="n">erosion</span> <span class="o">=</span> <span class="mi">1</span> <span class="o">-</span> <span class="n">torch</span><span class="o">.</span><span class="n">clamp</span><span class="p">(</span>
        <span class="n">F</span><span class="o">.</span><span class="n">conv2d</span><span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">padded_x</span><span class="p">,</span> <span class="n">_kernel</span><span class="p">,</span> <span class="n">groups</span><span class="o">=</span><span class="n">x</span><span class="o">.</span><span class="n">size</span><span class="p">(</span><span class="mi">1</span><span class="p">)),</span>
        <span class="mi">0</span><span class="p">,</span> <span class="mi">1</span>
    <span class="p">)</span>
    <span class="c1"># Finally the edge features are the result of subtracting dilation by erosion.</span>
    <span class="c1"># i.e for a simple 1D one-hot input:    [0, 0, 0, 1, 1, 1, 0, 0, 0], using sliding kernel with size 3: [1, 1, 1]</span>
    <span class="c1"># Dilated features:                     [0, 0, 1, 1, 1, 1, 1, 0, 0]</span>
    <span class="c1"># Erosed inverse features:              [0, 0, 0, 0, 1, 0, 0, 0, 0]</span>
    <span class="c1"># Edge features: dilation - erosion:    [0, 0, 1, 1, 0, 1, 1, 0, 0]</span>
    <span class="n">edge</span> <span class="o">=</span> <span class="n">dilation</span> <span class="o">-</span> <span class="n">erosion</span>
    <span class="k">if</span> <span class="n">flatten_channels</span><span class="p">:</span>
        <span class="c1"># use max operator across channels. Equivalent to logical or for input with binary values [0, 1].</span>
        <span class="n">edge</span> <span class="o">=</span> <span class="n">edge</span><span class="o">.</span><span class="n">max</span><span class="p">(</span><span class="n">dim</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">keepdim</span><span class="o">=</span><span class="kc">True</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
    <span class="k">return</span> <span class="n">edge</span></div>
<div class="viewcode-block" id="target_to_binary_edge"><a class="viewcode-back" href="../../../../super_gradients.training.utils.html#super_gradients.training.utils.segmentation_utils.target_to_binary_edge">[docs]</a><span class="k">def</span> <span class="nf">target_to_binary_edge</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">kernel_size</span><span class="p">:</span> <span class="nb">int</span><span class="p">,</span>
                          <span class="n">ignore_index</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="n">flatten_channels</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="o">-&gt;</span> <span class="n">torch</span><span class="o">.</span><span class="n">Tensor</span><span class="p">:</span>
    <span class="sd">&quot;&quot;&quot;</span>
<span class="sd">    Utils function to create edge feature maps from target.</span>
<span class="sd">    :param target: Class labels long tensor, with shape [N, H, W]</span>
<span class="sd">    :param num_classes: num of classes in datasets excluding ignore label, this is the output channels of the one hot</span>
<span class="sd">        result.</span>
<span class="sd">    :param kernel_size: kernel size of dilation erosion convolutions. The result edge widths depends on this argument as</span>
<span class="sd">        follows: `edge_width = kernel - 1`</span>
<span class="sd">    :param flatten_channels: Whether to apply logical or across channels dimension, if at least one pixel class is</span>
<span class="sd">        considered as edge pixel flatten value is 1. If set as `False` the output tensor shape is [B, C, H, W], else</span>
<span class="sd">        [B, 1, H, W]. Default is `True`.</span>
<span class="sd">    :return: one_hot edge torch.Tensor.</span>
<span class="sd">    &quot;&quot;&quot;</span>
    <span class="n">one_hot</span> <span class="o">=</span> <span class="n">to_one_hot</span><span class="p">(</span><span class="n">target</span><span class="p">,</span> <span class="n">num_classes</span><span class="o">=</span><span class="n">num_classes</span><span class="p">,</span> <span class="n">ignore_index</span><span class="o">=</span><span class="n">ignore_index</span><span class="p">)</span>
    <span class="k">return</span> <span class="n">one_hot_to_binary_edge</span><span class="p">(</span><span class="n">one_hot</span><span class="p">,</span> <span class="n">kernel_size</span><span class="o">=</span><span class="n">kernel_size</span><span class="p">,</span> <span class="n">flatten_channels</span><span class="o">=</span><span class="n">flatten_channels</span><span class="p">)</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>

          
        
      

  
Tip!
Press p or to see the previous file or, n or to see the next file
Deci-AI / super-gradients connected to https://github.com/Deci-AI/super-gradients.git

#396 Trainer constructor cleanup

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