:orphan:

Tutorials
=========

On this page, you will find tutorials and insights on TorchUncertainty. Don't
hesitate to open an issue if you have any question or suggestion for tutorials.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. thumbnail-parent-div-close

.. raw:: html

    </div>


Classification
--------------

Tutorials for modeling uncertainty in classification tasks.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In this tutorial, we will train a LeNet classifier on the MNIST dataset using TorchUncertainty. You will discover two of the core tools from TorchUncertainty, namely">

.. only:: html

  .. image:: /auto_tutorials/Classification/images/thumb/sphx_glr_tutorial_classification_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Classification_tutorial_classification.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Training a LeNet for Image Classification with TorchUncertainty</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In this tutorial, we explore how to assess a model&#x27;s robustness when faced with distribution shifts. Specifically, we will:">

.. only:: html

  .. image:: /auto_tutorials/Classification/images/thumb/sphx_glr_tutorial_distribution_shift_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Classification_tutorial_distribution_shift.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Evaluating Model Performance Under Distribution Shift with TorchUncertainty</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial aims to provide an introductory overview of Deep Evidential Classification (DEC) using a practical example. We demonstrate an application of DEC by tackling the toy-problem of fitting the MNIST dataset using a Multi-Layer Perceptron (MLP) neural network model. The output of the MLP is modeled as a Dirichlet distribution. The MLP is trained by minimizing the DEC loss function, composed of a Bayesian risk square error loss and a regularization term based on KL Divergence.">

.. only:: html

  .. image:: /auto_tutorials/Classification/images/thumb/sphx_glr_tutorial_evidential_classification_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Classification_tutorial_evidential_classification.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Deep Evidential Classification on a Toy Example</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial demonstrates how to perform OOD detection using TorchUncertainty&#x27;s ClassificationRoutine with a ResNet18 model trained on CIFAR-10, evaluating its performance with SVHN as the OOD dataset.">

.. only:: html

  .. image:: /auto_tutorials/Classification/images/thumb/sphx_glr_tutorial_ood_detection_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Classification_tutorial_ood_detection.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Out-of-distribution detection with TorchUncertainty</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial is heavily inspired by PyTorch&#x27;s Training a Classifier tutorial.">

.. only:: html

  .. image:: /auto_tutorials/Classification/images/thumb/sphx_glr_tutorial_pe_cifar10_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Classification_tutorial_pe_cifar10.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">From a Standard Classifier to a Packed-Ensemble</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


Regression with Uncertainty
---------------------------

Tutorials for modeling predictive uncertainty in regression tasks.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial provides an introduction to probabilistic regression in TorchUncertainty.">

.. only:: html

  .. image:: /auto_tutorials/Regression/images/thumb/sphx_glr_tutorial_der_cubic_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Regression_tutorial_der_cubic.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Deep Evidential Regression on a Toy Example</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial aims to provide an overview of some utilities in TorchUncertainty for probabilistic regression. Contrary to pointwise prediction, probabilistic regression consists - in TorchUncertainty&#x27;s context - in predicting the parameters of a predefined distribution that fit best some training dataset. The distribution&#x27;s formulation is fixed but the parameters are different for all data points, we say that the distribution is heteroscedastic.">

.. only:: html

  .. image:: /auto_tutorials/Regression/images/thumb/sphx_glr_tutorial_probabilistic_regression_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Regression_tutorial_probabilistic_regression.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Deep Probabilistic Regression</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In this tutorial, we will train a multi-layer-perceptron on a UCI regression dataset using TorchUncertainty. You will discover two of the core tools from TorchUncertainty, namely">

.. only:: html

  .. image:: /auto_tutorials/Regression/images/thumb/sphx_glr_tutorial_regression_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Regression_tutorial_regression.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Training an MLP for Tabular Regression with TorchUncertainty</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


Post-hoc Methods
----------------

Tutorials focused on improving model with post-hoc techniques.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="We evaluate the model&#x27;s performance both before and after applying different conformal predictors (THR, APS, RAPS), and visualize how conformal prediction estimates the prediction sets.">

.. only:: html

  .. image:: /auto_tutorials/Post_Hoc_Methods/images/thumb/sphx_glr_tutorial_conformal_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Post_Hoc_Methods_tutorial_conformal.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Conformal Prediction on CIFAR-10 with TorchUncertainty</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In this tutorial, we use TorchUncertainty to improve the calibration of the top-label predictions and the reliability of the underlying neural network.">

.. only:: html

  .. image:: /auto_tutorials/Post_Hoc_Methods/images/thumb/sphx_glr_tutorial_scaler_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Post_Hoc_Methods_tutorial_scaler.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Improve Top-label Calibration with Temperature Scaling</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


Bayesian Methods
-----------------

Tutorials for Bayesian approaches to uncertainty estimation.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In this tutorial, we will train a variational inference Bayesian Neural Network (viBNN) LeNet classifier on the MNIST dataset.">

.. only:: html

  .. image:: /auto_tutorials/Bayesian_Methods/images/thumb/sphx_glr_tutorial_bayesian_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Bayesian_Methods_tutorial_bayesian.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Training a Bayesian Neural Network in 20 seconds</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In this tutorial, we train a LeNet classifier on the MNIST dataset using Monte-Carlo Batch Normalization (MCBN), a post-hoc Bayesian approximation method.">

.. only:: html

  .. image:: /auto_tutorials/Bayesian_Methods/images/thumb/sphx_glr_tutorial_mc_batch_norm_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Bayesian_Methods_tutorial_mc_batch_norm.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Training a LeNet with Monte Carlo Batch Normalization</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="In this tutorial, we will train a LeNet classifier on the MNIST dataset using Monte-Carlo Dropout (MC Dropout), a computationally efficient Bayesian approximation method. To estimate the predictive mean and uncertainty (variance), we perform multiple forward passes through the network with dropout layers enabled in train mode.">

.. only:: html

  .. image:: /auto_tutorials/Bayesian_Methods/images/thumb/sphx_glr_tutorial_mc_dropout_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Bayesian_Methods_tutorial_mc_dropout.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Training a LeNet with Monte-Carlo Dropout</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial demonstrates how to train a segmentation model on the MUAD dataset using TorchUncertainty. MUAD is a synthetic dataset designed for evaluating autonomous driving under diverse uncertainties. It includes 10,413 images across training, validation, and test sets, featuring adverse weather, lighting conditions, and out-of-distribution (OOD) objects. The dataset supports tasks like semantic segmentation, depth estimation, and object detection.">

.. only:: html

  .. image:: /auto_tutorials/Bayesian_Methods/images/thumb/sphx_glr_tutorial_muad_mc_drop_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Bayesian_Methods_tutorial_muad_mc_drop.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Monte Carlo Dropout for Semantic Segmentation on MUAD</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


Ensemble Methods
-----------------

Tutorials for ensemble-based techniques to improve uncertainty estimation.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial is adapted from a notebook part of a lecture given at the Helmholtz AI Conference by Sebastian Starke, Peter Steinbach, Gianni Franchi, and Olivier Laurent.">

.. only:: html

  .. image:: /auto_tutorials/Ensemble_Methods/images/thumb/sphx_glr_tutorial_from_de_to_pe_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Ensemble_Methods_tutorial_from_de_to_pe.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Improved Ensemble parameter-efficiency with Packed-Ensembles</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


Segmentation
---------------------------

Tutorials for modeling uncertainty in Segmentation tasks.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial demonstrates how to train a segmentation model on the MUAD dataset using TorchUncertainty. MUAD is a synthetic dataset designed for evaluating autonomous driving under diverse uncertainties. It includes 10,413 images across training, validation, and test sets, featuring adverse weather, lighting conditions, and out-of-distribution (OOD) objects. The dataset supports tasks like semantic segmentation, depth estimation, and object detection.">

.. only:: html

  .. image:: /auto_tutorials/Segmentation/images/thumb/sphx_glr_tutorial_muad_deep_en_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Segmentation_tutorial_muad_deep_en.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Deep ensembles Segmentation Tutorial using Muad Dataset</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial demonstrates how to train a segmentation model on the MUAD dataset using TorchUncertainty. MUAD is a synthetic dataset designed for evaluating autonomous driving under diverse uncertainties. It includes 10,413 images across training, validation, and test sets, featuring adverse weather, lighting conditions, and out-of-distribution (OOD) objects. The dataset supports tasks like semantic segmentation, depth estimation, and object detection.">

.. only:: html

  .. image:: /auto_tutorials/Segmentation/images/thumb/sphx_glr_tutorial_muad_packed_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Segmentation_tutorial_muad_packed.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Packed ensembles Segmentation Tutorial using Muad Dataset</div>
    </div>


.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial demonstrates how to train a segmentation model on the MUAD dataset using TorchUncertainty. MUAD is a synthetic dataset designed for evaluating autonomous driving under diverse uncertainties. It includes 10,413 images across training, validation, and test sets, featuring adverse weather, lighting conditions, and out-of-distribution (OOD) objects. The dataset supports tasks like semantic segmentation, depth estimation, and object detection.">

.. only:: html

  .. image:: /auto_tutorials/Segmentation/images/thumb/sphx_glr_tutorial_muad_seg_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Segmentation_tutorial_muad_seg.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Segmentation Tutorial using Muad Dataset</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


Data Augmentation
-----------------

Tutorials illustrating data augmentation functionnalities in Torch-Uncertainty.



.. raw:: html

    <div class="sphx-glr-thumbnails">

.. thumbnail-parent-div-open

.. raw:: html

    <div class="sphx-glr-thumbcontainer" tooltip="This tutorial shows the impact of the different corruption transforms available in the TorchUncertainty library. These corruption transforms were first proposed in the paper Benchmarking Neural Network Robustness to Common Corruptions and Perturbations by Dan Hendrycks and Thomas Dietterich.">

.. only:: html

  .. image:: /auto_tutorials/Data_Augmentation/images/thumb/sphx_glr_tutorial_corruption_thumb.png
    :alt:

  :ref:`sphx_glr_auto_tutorials_Data_Augmentation_tutorial_corruption.py`

.. raw:: html

      <div class="sphx-glr-thumbnail-title">Corrupting Images with TorchUncertainty to Benchmark Robustness</div>
    </div>


.. thumbnail-parent-div-close

.. raw:: html

    </div>


.. toctree::
   :hidden:
   :includehidden:


   /auto_tutorials/Classification/index.rst
   /auto_tutorials/Regression/index.rst
   /auto_tutorials/Post_Hoc_Methods/index.rst
   /auto_tutorials/Bayesian_Methods/index.rst
   /auto_tutorials/Ensemble_Methods/index.rst
   /auto_tutorials/Segmentation/index.rst
   /auto_tutorials/Data_Augmentation/index.rst


.. only:: html

  .. container:: sphx-glr-footer sphx-glr-footer-gallery

    .. container:: sphx-glr-download sphx-glr-download-python

      :download:`Download all examples in Python source code: auto_tutorials_python.zip </auto_tutorials/auto_tutorials_python.zip>`

    .. container:: sphx-glr-download sphx-glr-download-jupyter

      :download:`Download all examples in Jupyter notebooks: auto_tutorials_jupyter.zip </auto_tutorials/auto_tutorials_jupyter.zip>`


.. only:: html

 .. rst-class:: sphx-glr-signature

    `Gallery generated by Sphinx-Gallery <https://sphinx-gallery.github.io>`_