@@ -38,6 +38,7 @@
         </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>
@@ -120,7 +121,7 @@ The type of accuracy to calculate, e.g. topk=(1,5) returns accuracy for top-1 an
 <dt class="sig sig-object py" id="super_gradients.training.metrics.classification_metrics.Accuracy.update">
 <span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="s
 <dd><p>Update state with predictions and targets. See
-<span class="xref std std-ref">references/modules:input types</span> for more information on input
+<span class="xref std std-ref">pages/classification:input types</span> for more information on input
 types.</p>
 <dl class="field-list simple">
 <dt class="field-odd">Parameters</dt>
@@ -142,6 +143,16 @@ types.</p>
 <span class="sig-name descname"><span class="pre">total</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.classification_metrics.Accuracy.total" title="Permalink to this definition"></a></dt>
 <dd></dd></dl>
 
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.classification_metrics.Accuracy.mode">
+<span class="sig-name descname"><span class="pre">mode</span></span><em class="property"><span class="pre">:</span> <span class="pre">DataType</span></em><a class="headerlink" href="#super_gradients.training.metrics.classification_metrics.Accuracy.mode" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.classification_metrics.Accuracy.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.classification_metrics.Accuracy.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
 </dd></dl>
 
 <dl class="py class">
@@ -186,64 +197,95 @@ distributed backend.</p>
 </section>
 <section id="module-super_gradients.training.metrics.detection_metrics">
 <span id="super-gradients-training-metrics-detection-metrics-module"></span><h2>super_gradients.training.metrics.detection_metrics module<a class="headerlink" href="#module-super_gradients.training.metrics.detection_metrics" title="Permalink to this headline"></a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.compute_ap">
-<span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.detection_metrics.</span></span><span class="sig-name descname"><span class="pre">compute_ap</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">recall</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">precision</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">method</span></span><span class="p"><span cla
-<dd><p>Compute the average precision, given the recall and precision curves.
-Source: <a class="reference external" href="https://github.com/rbgirshick/py-faster-rcnn">https://github.com/rbgirshick/py-faster-rcnn</a>.
-# Arguments</p>
-<blockquote>
-<div><dl class="field-list simple">
-<dt class="field-odd">param recall</dt>
-<dd class="field-odd"><p>The recall curve - ndarray [1, points in curve]</p>
-</dd>
-<dt class="field-even">param precision</dt>
-<dd class="field-even"><p>The precision curve - ndarray [1, points in curve]</p>
-</dd>
-<dt class="field-odd">param method</dt>
-<dd class="field-odd"><p>‘continuous’, ‘interp’</p>
-</dd>
-</dl>
-</div></blockquote>
-<dl class="simple">
-<dt># Returns</dt><dd><p>The average precision as computed in py-faster-rcnn.</p>
-</dd>
-</dl>
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.detection_metrics.</span></span><span class="sig-name descname"><span class="pre">DetectionMetrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">num_cls:</span> <span class="pre">int</span></em>, <em class="sig-param"><span class="pre">post_prediction_callback:</span> <span class="pre">Optional[super_gradients.
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.metric.Metric</span></code></p>
+<p>Metric class for computing F1, Precision, Recall and Mean Average Precision.</p>
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.num_cls">
+<span class="sig-name descname"><span class="pre">num_cls</span></span><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.num_cls" title="Permalink to this definition"></a></dt>
+<dd><p>Number of classes.</p>
 </dd></dl>
 
-<dl class="py function">
-<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.ap_per_class">
-<span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.detection_metrics.</span></span><span class="sig-name descname"><span class="pre">ap_per_class</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">tp</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">conf</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">pred_cls</span></span></em>, <em class="sig-param"><
-<dd><p>Compute the average precision, given the recall and precision curves.
-Source: <a class="reference external" href="https://github.com/rafaelpadilla/Object-Detection-Metrics">https://github.com/rafaelpadilla/Object-Detection-Metrics</a>.
-# Arguments</p>
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.post_prediction_callback">
+<span class="sig-name descname"><span class="pre">post_prediction_callback</span></span><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.post_prediction_callback" title="Permalink to this definition"></a></dt>
+<dd><p>DetectionPostPredictionCallback to be applied on net’s output prior
+to the metric computation (NMS).</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.normalize_targets">
+<span class="sig-name descname"><span class="pre">normalize_targets</span></span><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.normalize_targets" title="Permalink to this definition"></a></dt>
+<dd><p>Whether to normalize bbox coordinates by image size (default=False).</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.iou_thresholds">
+<span class="sig-name descname"><span class="pre">iou_thresholds</span></span><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.iou_thresholds" title="Permalink to this definition"></a></dt>
+<dd><p>IoU threshold to compute the mAP (default=torch.linspace(0.5, 0.95, 10)).</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.recall_thresholds">
+<span class="sig-name descname"><span class="pre">recall_thresholds</span></span><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.recall_thresholds" title="Permalink to this definition"></a></dt>
+<dd><p>Recall threshold to compute the mAP (default=torch.linspace(0, 1, 101)).</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.score_threshold">
+<span class="sig-name descname"><span class="pre">score_threshold</span></span><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.score_threshold" title="Permalink to this definition"></a></dt>
+<dd><p>Score threshold to compute Recall, Precision and F1 (default=0.1)</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.top_k_predictions">
+<span class="sig-name descname"><span class="pre">top_k_predictions</span></span><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.top_k_predictions" title="Permalink to this definition"></a></dt>
+<dd><p>Number of predictions per class used to compute metrics, ordered by confidence score
+(default=100)</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.dist_sync_on_step">
+<span class="sig-name descname"><span class="pre">dist_sync_on_step</span></span><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.dist_sync_on_step" title="Permalink to this definition"></a></dt>
+<dd><p>Synchronize metric state across processes at each <code class="docutils literal notranslate"><span class="pre">forward()</span></code>
+before returning the value at the step. (default=False)</p>
 <blockquote>
-<div><p>tp:    True positives (nparray, nx1 or nx10).
-conf:  Objectness value from 0-1 (nparray).
-pred_cls: Predicted object classes (nparray).
-target_cls: True object classes (nparray).</p>
-</div></blockquote>
-<dl class="simple">
-<dt># Returns</dt><dd><p>The average precision as computed in py-faster-rcnn.</p>
+<div><dl class="simple">
+<dt>accumulate_on_cpu:     Run on CPU regardless of device used in other parts.</dt><dd><p>This is to avoid “CUDA out of memory” that might happen on GPU (default False)</p>
 </dd>
 </dl>
+</div></blockquote>
 </dd></dl>
 
-<dl class="py class">
-<dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics">
-<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.detection_metrics.</span></span><span class="sig-name descname"><span class="pre">DetectionMetrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">num_cls</span></em>, <em class="sig-param"><span class="pre">post_prediction_callback:</span> <span class="pre">Optional[super_gradients.training.utils.detection_utils
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.metric.Metric</span></code></p>
 <dl class="py method">
 <dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.update">
-<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="s
-<dd><p>Override this method to update the state variables of your metric class.</p>
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="p"><span class="pre">:</span></sp
+<dd><p>Apply NMS and match all the predictions and targets of a given batch, and update the metric state accordingly.</p>
+<dl class="simple">
+<dt>:param preds<span class="classifier">Raw output of the model, the format might change from one model to another, but has to fit</span></dt><dd><p>the input format of the post_prediction_callback</p>
+</dd>
+</dl>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>target</strong> – Targets for all images of shape (total_num_targets, 6)
+format:  (index, x, y, w, h, label) where x,y,w,h are in range [0,1]</p></li>
+<li><p><strong>device</strong> – Device to run on</p></li>
+<li><p><strong>inputs</strong> – Input image tensor of shape (batch_size, n_img, height, width)</p></li>
+<li><p><strong>crowd_targets</strong> – Crowd targets for all images of shape (total_num_targets, 6)
+format:  (index, x, y, w, h, label) where x,y,w,h are in range [0,1]</p></li>
+</ul>
+</dd>
+</dl>
 </dd></dl>
 
 <dl class="py method">
 <dt class="sig sig-object py" id="super_gradients.training.metrics.detection_metrics.DetectionMetrics.compute">
-<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/detection_metrics.html#DetectionMetrics.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.detection_metrics.DetectionMetrics.compute" title="Permalink to this definition"></a></dt>
-<dd><p>Override this method to compute the final metric value from state variables synchronized across the
-distributed backend.</p>
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> &#x2192; <span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">,</span> </span><span class="pre">torch.Tensor</span><s
+<dd><p>Compute the metrics for all the accumulated results.
+:return: Metrics of interest</p>
 </dd></dl>
 
 </dd></dl>
@@ -251,11 +293,6 @@ distributed backend.</p>
 </section>
 <section id="module-super_gradients.training.metrics.metric_utils">
 <span id="super-gradients-training-metrics-metric-utils-module"></span><h2>super_gradients.training.metrics.metric_utils module<a class="headerlink" href="#module-super_gradients.training.metrics.metric_utils" title="Permalink to this headline"></a></h2>
-<dl class="py function">
-<dt class="sig sig-object py" id="super_gradients.training.metrics.metric_utils.calc_batch_prediction_detection_metrics_per_class">
-<span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.metric_utils.</span></span><span class="sig-name descname"><span class="pre">calc_batch_prediction_detection_metrics_per_class</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">metrics</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dataset_interface</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">io
-<dd></dd></dl>
-
 <dl class="py function">
 <dt class="sig sig-object py" id="super_gradients.training.metrics.metric_utils.get_logging_values">
 <span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.metric_utils.</span></span><span class="sig-name descname"><span class="pre">get_logging_values</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">loss_loggings</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><a class="reference internal" href="super_gradients.training.utils.html#super_gradients.training.utils.utils.AverageMe
@@ -343,9 +380,24 @@ computed and returned.</p>
 <span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">intersection_and_union</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">im_pred</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">im_lab</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">num_class</span></span></em><spa
 <dd></dd></dl>
 
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.AbstractMetricsArgsPrepFn">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">AbstractMetricsArgsPrepFn</span></span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_metrics.html#AbstractMetricsArgsPrepFn"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" hr
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">abc.ABC</span></code></p>
+<p>Abstract preprocess metrics arguments class.</p>
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.PreprocessSegmentationMetricsArgs">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">PreprocessSegmentationMetricsArgs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">apply_arg_max</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <
+<dd><p>Bases: <a class="reference internal" href="#super_gradients.training.metrics.segmentation_metrics.AbstractMetricsArgsPrepFn" title="super_gradients.training.metrics.segmentation_metrics.AbstractMetricsArgsPrepFn"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.metrics.segmentation_metrics.AbstractMetricsArgsPrepFn</span></code></a></p>
+<p>Default segmentation inputs preprocess function before updating segmentation metrics, handles multiple inputs and
+apply normalizations.</p>
+</dd></dl>
+
 <dl class="py class">
 <dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.PixelAccuracy">
-<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">PixelAccuracy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ignore_label</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-</span> <span class="pre">100
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">PixelAccuracy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ignore_label</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-</span> <span class="pre">100
 <dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.metric.Metric</span></code></p>
 <dl class="py method">
 <dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.PixelAccuracy.update">
@@ -364,8 +416,8 @@ distributed backend.</p>
 
 <dl class="py class">
 <dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.IoU">
-<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">IoU</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">num_classes</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span>
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.classification.iou.IoU</span></code></p>
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">IoU</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">num_classes</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span 
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.classification.jaccard.JaccardIndex</span></code></p>
 <dl class="py method">
 <dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.IoU.update">
 <span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_m
@@ -385,15 +437,20 @@ distributed backend.</p>
 <span class="sig-name descname"><span class="pre">confmat</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.IoU.confmat" title="Permalink to this definition"></a></dt>
 <dd></dd></dl>
 
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.IoU.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.IoU.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
 </dd></dl>
 
 <dl class="py class">
-<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryIOU">
-<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">BinaryIOU</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em 
-<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.classification.iou.IoU</span></code></p>
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.Dice">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">Dice</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">num_classes</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.classification.jaccard.JaccardIndex</span></code></p>
 <dl class="py method">
-<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryIOU.update">
-<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_m
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.Dice.update">
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_m
 <dd><p>Update state with predictions and targets.</p>
 <dl class="field-list simple">
 <dt class="field-odd">Parameters</dt>
@@ -405,6 +462,28 @@ distributed backend.</p>
 </dl>
 </dd></dl>
 
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.Dice.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> &#x2192; <span class="pre">torch.Tensor</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_metrics.html#Dice.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.Dice.compute" title="Permalink to this defin
+<dd><p>Computes Dice coefficient</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.Dice.confmat">
+<span class="sig-name descname"><span class="pre">confmat</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.Dice.confmat" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.Dice.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.Dice.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryIOU">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">BinaryIOU</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em 
+<dd><p>Bases: <a class="reference internal" href="#super_gradients.training.metrics.segmentation_metrics.IoU" title="super_gradients.training.metrics.segmentation_metrics.IoU"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.metrics.segmentation_metrics.IoU</span></code></a></p>
 <dl class="py method">
 <dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryIOU.compute">
 <span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_metrics.html#BinaryIOU.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.BinaryIOU.compute" title="Permalink to this definition"></a></dt>
@@ -416,11 +495,369 @@ distributed backend.</p>
 <span class="sig-name descname"><span class="pre">confmat</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.BinaryIOU.confmat" title="Permalink to this definition"></a></dt>
 <dd></dd></dl>
 
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryIOU.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.BinaryIOU.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryDice">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.segmentation_metrics.</span></span><span class="sig-name descname"><span class="pre">BinaryDice</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em
+<dd><p>Bases: <a class="reference internal" href="#super_gradients.training.metrics.segmentation_metrics.Dice" title="super_gradients.training.metrics.segmentation_metrics.Dice"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.metrics.segmentation_metrics.Dice</span></code></a></p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryDice.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_metrics.html#BinaryDice.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.BinaryDice.compute" title="Permalink to this definition"></a></dt>
+<dd><p>Computes Dice coefficient</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryDice.confmat">
+<span class="sig-name descname"><span class="pre">confmat</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.BinaryDice.confmat" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.segmentation_metrics.BinaryDice.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.segmentation_metrics.BinaryDice.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
 </dd></dl>
 
 </section>
 <section id="module-super_gradients.training.metrics">
 <span id="module-contents"></span><h2>Module contents<a class="headerlink" href="#module-super_gradients.training.metrics" title="Permalink to this headline"></a></h2>
+<dl class="py function">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.accuracy">
+<span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">accuracy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">output</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">topk</span></span><span class="o"><span class="pre">=</span></span><
+<dd><p>Computes the precision&#64;k for the specified values of k
+:param output: Tensor / Numpy / List</p>
+<blockquote>
+<div><p>The prediction</p>
+</div></blockquote>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>target</strong> – Tensor / Numpy / List
+The corresponding lables</p></li>
+<li><p><strong>topk</strong> – tuple
+The type of accuracy to calculate, e.g. topk=(1,5) returns accuracy for top-1 and top-5</p></li>
+</ul>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Accuracy">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">Accuracy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.classification.accuracy.Accuracy</span></code></p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Accuracy.update">
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="s
+<dd><p>Update state with predictions and targets. See
+<span class="xref std std-ref">pages/classification:input types</span> for more information on input
+types.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>preds</strong> – Predictions from model (logits, probabilities, or labels)</p></li>
+<li><p><strong>target</strong> – Ground truth labels</p></li>
+</ul>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Accuracy.correct">
+<span class="sig-name descname"><span class="pre">correct</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.Accuracy.correct" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Accuracy.total">
+<span class="sig-name descname"><span class="pre">total</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.Accuracy.total" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Accuracy.mode">
+<span class="sig-name descname"><span class="pre">mode</span></span><em class="property"><span class="pre">:</span> <span class="pre">DataType</span></em><a class="headerlink" href="#super_gradients.training.metrics.Accuracy.mode" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Accuracy.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.Accuracy.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Top5">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">Top5</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span class="sig-paren">)</span
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.metric.Metric</span></code></p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Top5.update">
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="s
+<dd><p>Override this method to update the state variables of your metric class.</p>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Top5.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/classification_metrics.html#Top5.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.Top5.compute" title="Permalink to this definition"></a></dt>
+<dd><p>Override this method to compute the final metric value from state variables synchronized across the
+distributed backend.</p>
+</dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.ToyTestClassificationMetric">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">ToyTestClassificationMetric</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">False</span></span></em><span cl
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.metric.Metric</span></code></p>
+<p>Dummy classification Mettric object returning 0 always (for testing).</p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.ToyTestClassificationMetric.update">
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="s
+<dd><p>Override this method to update the state variables of your metric class.</p>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.ToyTestClassificationMetric.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/classification_metrics.html#ToyTestClassificationMetric.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.ToyTestClassificationMetric.compute" title="Permalink to this definition"></a></dt>
+<dd><p>Override this method to compute the final metric value from state variables synchronized across the
+distributed backend.</p>
+</dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">DetectionMetrics</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="pre">num_cls:</span> <span class="pre">int</span></em>, <em class="sig-param"><span class="pre">post_prediction_callback:</span> <span class="pre">Optional[super_gradients.training.utils.det
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.metric.Metric</span></code></p>
+<p>Metric class for computing F1, Precision, Recall and Mean Average Precision.</p>
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.num_cls">
+<span class="sig-name descname"><span class="pre">num_cls</span></span><a class="headerlink" href="#super_gradients.training.metrics.DetectionMetrics.num_cls" title="Permalink to this definition"></a></dt>
+<dd><p>Number of classes.</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.post_prediction_callback">
+<span class="sig-name descname"><span class="pre">post_prediction_callback</span></span><a class="headerlink" href="#super_gradients.training.metrics.DetectionMetrics.post_prediction_callback" title="Permalink to this definition"></a></dt>
+<dd><p>DetectionPostPredictionCallback to be applied on net’s output prior
+to the metric computation (NMS).</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.normalize_targets">
+<span class="sig-name descname"><span class="pre">normalize_targets</span></span><a class="headerlink" href="#super_gradients.training.metrics.DetectionMetrics.normalize_targets" title="Permalink to this definition"></a></dt>
+<dd><p>Whether to normalize bbox coordinates by image size (default=False).</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.iou_thresholds">
+<span class="sig-name descname"><span class="pre">iou_thresholds</span></span><a class="headerlink" href="#super_gradients.training.metrics.DetectionMetrics.iou_thresholds" title="Permalink to this definition"></a></dt>
+<dd><p>IoU threshold to compute the mAP (default=torch.linspace(0.5, 0.95, 10)).</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.recall_thresholds">
+<span class="sig-name descname"><span class="pre">recall_thresholds</span></span><a class="headerlink" href="#super_gradients.training.metrics.DetectionMetrics.recall_thresholds" title="Permalink to this definition"></a></dt>
+<dd><p>Recall threshold to compute the mAP (default=torch.linspace(0, 1, 101)).</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.score_threshold">
+<span class="sig-name descname"><span class="pre">score_threshold</span></span><a class="headerlink" href="#super_gradients.training.metrics.DetectionMetrics.score_threshold" title="Permalink to this definition"></a></dt>
+<dd><p>Score threshold to compute Recall, Precision and F1 (default=0.1)</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.top_k_predictions">
+<span class="sig-name descname"><span class="pre">top_k_predictions</span></span><a class="headerlink" href="#super_gradients.training.metrics.DetectionMetrics.top_k_predictions" title="Permalink to this definition"></a></dt>
+<dd><p>Number of predictions per class used to compute metrics, ordered by confidence score
+(default=100)</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.dist_sync_on_step">
+<span class="sig-name descname"><span class="pre">dist_sync_on_step</span></span><a class="headerlink" href="#super_gradients.training.metrics.DetectionMetrics.dist_sync_on_step" title="Permalink to this definition"></a></dt>
+<dd><p>Synchronize metric state across processes at each <code class="docutils literal notranslate"><span class="pre">forward()</span></code>
+before returning the value at the step. (default=False)</p>
+<blockquote>
+<div><dl class="simple">
+<dt>accumulate_on_cpu:     Run on CPU regardless of device used in other parts.</dt><dd><p>This is to avoid “CUDA out of memory” that might happen on GPU (default False)</p>
+</dd>
+</dl>
+</div></blockquote>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.update">
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">device</span></span><span class="p"><span class="pre">:</span></sp
+<dd><p>Apply NMS and match all the predictions and targets of a given batch, and update the metric state accordingly.</p>
+<dl class="simple">
+<dt>:param preds<span class="classifier">Raw output of the model, the format might change from one model to another, but has to fit</span></dt><dd><p>the input format of the post_prediction_callback</p>
+</dd>
+</dl>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>target</strong> – Targets for all images of shape (total_num_targets, 6)
+format:  (index, x, y, w, h, label) where x,y,w,h are in range [0,1]</p></li>
+<li><p><strong>device</strong> – Device to run on</p></li>
+<li><p><strong>inputs</strong> – Input image tensor of shape (batch_size, n_img, height, width)</p></li>
+<li><p><strong>crowd_targets</strong> – Crowd targets for all images of shape (total_num_targets, 6)
+format:  (index, x, y, w, h, label) where x,y,w,h are in range [0,1]</p></li>
+</ul>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.DetectionMetrics.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> &#x2192; <span class="pre">Dict</span><span class="p"><span class="pre">[</span></span><span class="pre">str</span><span class="p"><span class="pre">,</span> </span><span class="pre">Union</span><span class="p"><span class="pre">[</span></span><span class="pre">float</span><span class="p"><span class="pre">,</span> </span><span class="pre">torch.Tensor</span><s
+<dd><p>Compute the metrics for all the accumulated results.
+:return: Metrics of interest</p>
+</dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.PreprocessSegmentationMetricsArgs">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">PreprocessSegmentationMetricsArgs</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">apply_arg_max</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">bool</span></span> <span class="o"><span 
+<dd><p>Bases: <a class="reference internal" href="#super_gradients.training.metrics.segmentation_metrics.AbstractMetricsArgsPrepFn" title="super_gradients.training.metrics.segmentation_metrics.AbstractMetricsArgsPrepFn"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.metrics.segmentation_metrics.AbstractMetricsArgsPrepFn</span></code></a></p>
+<p>Default segmentation inputs preprocess function before updating segmentation metrics, handles multiple inputs and
+apply normalizations.</p>
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.PixelAccuracy">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">PixelAccuracy</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">ignore_label</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">-</span> <span class="pre">100</span></span></em>, 
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.metric.Metric</span></code></p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.PixelAccuracy.update">
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="s
+<dd><p>Override this method to update the state variables of your metric class.</p>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.PixelAccuracy.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_metrics.html#PixelAccuracy.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.PixelAccuracy.compute" title="Permalink to this definition"></a></dt>
+<dd><p>Override this method to compute the final metric value from state variables synchronized across the
+distributed backend.</p>
+</dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.IoU">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">IoU</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">num_classes</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span class
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.classification.jaccard.JaccardIndex</span></code></p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.IoU.update">
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_m
+<dd><p>Update state with predictions and targets.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>preds</strong> – Predictions from model</p></li>
+<li><p><strong>target</strong> – Ground truth values</p></li>
+</ul>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.IoU.confmat">
+<span class="sig-name descname"><span class="pre">confmat</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.IoU.confmat" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.IoU.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.IoU.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Dice">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">Dice</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">num_classes</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">int</span></span></em>, <em class="sig-param"><span class="n"><span clas
+<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">torchmetrics.classification.jaccard.JaccardIndex</span></code></p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Dice.update">
+<span class="sig-name descname"><span class="pre">update</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">preds</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">target</span></span><span class="p"><span class="pre">:</span></span> <span class="n"><span class="pre">torch.Tensor</span></span></em><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_m
+<dd><p>Update state with predictions and targets.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters</dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>preds</strong> – Predictions from model</p></li>
+<li><p><strong>target</strong> – Ground truth values</p></li>
+</ul>
+</dd>
+</dl>
+</dd></dl>
+
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Dice.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span> &#x2192; <span class="pre">torch.Tensor</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_metrics.html#Dice.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.Dice.compute" title="Permalink to this definition"></a></dt>
+<dd><p>Computes Dice coefficient</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Dice.confmat">
+<span class="sig-name descname"><span class="pre">confmat</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.Dice.confmat" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.Dice.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.Dice.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.BinaryIOU">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">BinaryIOU</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><sp
+<dd><p>Bases: <a class="reference internal" href="#super_gradients.training.metrics.segmentation_metrics.IoU" title="super_gradients.training.metrics.segmentation_metrics.IoU"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.metrics.segmentation_metrics.IoU</span></code></a></p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.BinaryIOU.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_metrics.html#BinaryIOU.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.BinaryIOU.compute" title="Permalink to this definition"></a></dt>
+<dd><p>Computes intersection over union (IoU)</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.BinaryIOU.confmat">
+<span class="sig-name descname"><span class="pre">confmat</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.BinaryIOU.confmat" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.BinaryIOU.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.BinaryIOU.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+</dd></dl>
+
+<dl class="py class">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.BinaryDice">
+<em class="property"><span class="pre">class</span> </em><span class="sig-prename descclassname"><span class="pre">super_gradients.training.metrics.</span></span><span class="sig-name descname"><span class="pre">BinaryDice</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">dist_sync_on_step</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">True</span></span></em>, <em class="sig-param"><s
+<dd><p>Bases: <a class="reference internal" href="#super_gradients.training.metrics.segmentation_metrics.Dice" title="super_gradients.training.metrics.segmentation_metrics.Dice"><code class="xref py py-class docutils literal notranslate"><span class="pre">super_gradients.training.metrics.segmentation_metrics.Dice</span></code></a></p>
+<dl class="py method">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.BinaryDice.compute">
+<span class="sig-name descname"><span class="pre">compute</span></span><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="_modules/super_gradients/training/metrics/segmentation_metrics.html#BinaryDice.compute"><span class="viewcode-link"><span class="pre">[source]</span></span></a><a class="headerlink" href="#super_gradients.training.metrics.BinaryDice.compute" title="Permalink to this definition"></a></dt>
+<dd><p>Computes Dice coefficient</p>
+</dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.BinaryDice.confmat">
+<span class="sig-name descname"><span class="pre">confmat</span></span><em class="property"><span class="pre">:</span> <span class="pre">torch.Tensor</span></em><a class="headerlink" href="#super_gradients.training.metrics.BinaryDice.confmat" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+<dl class="py attribute">
+<dt class="sig sig-object py" id="super_gradients.training.metrics.BinaryDice.training">
+<span class="sig-name descname"><span class="pre">training</span></span><em class="property"><span class="pre">:</span> <span class="pre">bool</span></em><a class="headerlink" href="#super_gradients.training.metrics.BinaryDice.training" title="Permalink to this definition"></a></dt>
+<dd></dd></dl>
+
+</dd></dl>
+
 </section>
 </section>