.. DO NOT EDIT. .. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY. .. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE: .. "checks_gallery/vision/distribution/plot_train_test_label_drift.py" .. LINE NUMBERS ARE GIVEN BELOW. .. only:: html .. note:: :class: sphx-glr-download-link-note Click :ref:`here ` to download the full example code .. rst-class:: sphx-glr-example-title .. _sphx_glr_checks_gallery_vision_distribution_plot_train_test_label_drift.py: Train Test Label Drift ********************** This notebooks provides an overview for using and understanding the vision label drift check. **Structure:** * `What is a label drift? <#what-is-a-label-drift>`__ * `Which Label Properties Are Used? <#which-label-properties-are-used>`__ * `Run check on a Classification task <#run-the-check-on-a-classification-task-mnist>`__ * `Run check on an Object Detection task <#run-the-check-on-an-object-detection-task-coco>`__ What is a label drift? ====================== The term drift (and all it's derivatives) is used to describe any change in the data compared to the data the model was trained on. Specifically, label drift indicates changes in the label we are trying to predict. Causes of label drift include: * Natural drift in the data, such as a certain class becoming more prevalent in the test set. For example, cronuts becoming more popular in a food classification dataset. * Labeling issues, such as an analyst drawing incorrect bounding boxes for an object detection task. How Does the TrainTestLabelDrift Check Work? ============================================ There are many methods to detect drift, that usually include statistical methods that aim to measure difference between 2 distributions. We experimented with various approaches and found that for detecting drift between 2 one-dimensional distributions, the following 2 methods give the best results: * For numerical features, the `Population Stability Index (PSI) `__ * For categorical features, the `Wasserstein Distance (Earth Mover's Distance) `__ However, one does not simply measure drift on a label, as they may be complex structures. These methods are implemented on label properties, as described in the next section. Using Label Properties to Detect Label Drift -------------------------------------------- In computer vision specifically, our labels may be complex, and measuring their drift is not a straightforward task. Therefore, we calculate drift on different properties of the labels, on which we can directly measure drift. Which Label Properties Are Used? ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ================ =================================== ========== Task Type Property name What is it ================ =================================== ========== Classification Samples Per Class Number of images per class Object Detection Samples Per Class Number of bounding boxes per class Object Detection Bounding Box Area Area of bounding box (height * width) Object Detection Number of Bounding Boxes Per Image Number of bounding box objects in each image ================ =================================== ========== Run the check on a Classification task (MNIST) ============================================== Imports ------- .. GENERATED FROM PYTHON SOURCE LINES 66-69 .. code-block:: default from deepchecks.vision.checks import TrainTestLabelDrift from deepchecks.vision.datasets.classification.mnist import load_dataset .. GENERATED FROM PYTHON SOURCE LINES 70-72 Loading Data ------------ .. GENERATED FROM PYTHON SOURCE LINES 72-77 .. code-block:: default train_ds = load_dataset(train=True, batch_size=64, object_type='VisionData') test_ds = load_dataset(train=False, batch_size=1000, object_type='VisionData') .. GENERATED FROM PYTHON SOURCE LINES 78-80 Running TrainTestLabelDrift on classification --------------------------------------------- .. GENERATED FROM PYTHON SOURCE LINES 80-84 .. code-block:: default check = TrainTestLabelDrift() check.run(train_ds, test_ds) .. rst-class:: sphx-glr-script-out Out: .. code-block:: none Validating Input: 0%| | 0/1 [00:00

Train Test Label Drift

Calculate label drift between train dataset and test dataset, using statistical measures.

Additional Outputs
The Drift score is a measure for the difference between two distributions. In this check, drift is measured for the distribution of the following label properties: ['Samples Per Class'].

Note - data sampling: Running on 10000 train data samples out of 60000. Sample size can be controlled with the "n_samples" parameter.



.. GENERATED FROM PYTHON SOURCE LINES 85-90 Understanding the results ------------------------- We can see there is almost no drift between the train & test labels. This means the split to train and test was good (as it is balanced and random). Let's check the performance of a simple model trained on MNIST. .. GENERATED FROM PYTHON SOURCE LINES 90-98 .. code-block:: default from deepchecks.vision.checks import ClassPerformance from deepchecks.vision.datasets.classification.mnist import \ load_model as load_mnist_model mnist_model = load_mnist_model(pretrained=True) ClassPerformance().run(train_ds, test_ds, mnist_model) .. rst-class:: sphx-glr-script-out Out: .. code-block:: none Validating Input: 0%| | 0/1 [00:00

Class Performance

Summarize given metrics on a dataset and model.

Additional Outputs

Note - data sampling: Running on 10000 train data samples out of 60000. Sample size can be controlled with the "n_samples" parameter.



.. GENERATED FROM PYTHON SOURCE LINES 99-105 MNIST with label drift ====================== Now, let's try to separate the MNIST dataset in a different manner that will result in a label drift, and see how it affects the performance. We are going to create a custom `collate_fn`` in the test dataset, that will select samples with class 0 in a 1/10 chances. .. GENERATED FROM PYTHON SOURCE LINES 105-112 .. code-block:: default import torch mnist_dataloader_train = load_dataset(train=True, batch_size=64, object_type='DataLoader') mnist_dataloader_test = load_dataset(train=False, batch_size=1000, object_type='DataLoader') full_mnist = torch.utils.data.ConcatDataset([mnist_dataloader_train.dataset, mnist_dataloader_test.dataset]) .. GENERATED FROM PYTHON SOURCE LINES 113-115 .. code-block:: default train_dataset, test_dataset = torch.utils.data.random_split(full_mnist, [60000,10000], generator=torch.Generator().manual_seed(42)) .. GENERATED FROM PYTHON SOURCE LINES 116-118 Inserting drift to the test set ------------------------------- .. GENERATED FROM PYTHON SOURCE LINES 118-143 .. code-block:: default import numpy as np from torch.utils.data._utils.collate import default_collate np.random.seed(42) def collate_test(batch): modified_batch = [] for item in batch: image, label = item if label == 0: if np.random.randint(5) == 0: modified_batch.append(item) else: modified_batch.append((image, 1)) else: modified_batch.append(item) return default_collate(modified_batch) mod_train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=64) mod_test_loader = torch.utils.data.DataLoader(test_dataset, batch_size=64, collate_fn=collate_test) .. GENERATED FROM PYTHON SOURCE LINES 144-150 .. code-block:: default from deepchecks.vision.datasets.classification.mnist import MNISTData mod_train_ds = MNISTData(mod_train_loader) mod_test_ds = MNISTData(mod_test_loader) .. GENERATED FROM PYTHON SOURCE LINES 151-153 Run the check ============= .. GENERATED FROM PYTHON SOURCE LINES 153-157 .. code-block:: default check = TrainTestLabelDrift() check.run(mod_train_ds, mod_test_ds) .. rst-class:: sphx-glr-script-out Out: .. code-block:: none Validating Input: 0%| | 0/1 [00:00

Train Test Label Drift

Calculate label drift between train dataset and test dataset, using statistical measures.

Additional Outputs
The Drift score is a measure for the difference between two distributions. In this check, drift is measured for the distribution of the following label properties: ['Samples Per Class'].

Note - data sampling: Running on 10000 train data samples out of 60000. Sample size can be controlled with the "n_samples" parameter.



.. GENERATED FROM PYTHON SOURCE LINES 158-162 Add a condition --------------- We could also add a condition to the check to alert us to changes in the label distribution, such as the one that occurred here. .. GENERATED FROM PYTHON SOURCE LINES 162-168 .. code-block:: default check = TrainTestLabelDrift().add_condition_drift_score_not_greater_than() check.run(mod_train_ds, mod_test_ds) # As we can see, the condition alerts us to the present of drift in the label. .. rst-class:: sphx-glr-script-out Out: .. code-block:: none Validating Input: 0%| | 0/1 [00:00

Train Test Label Drift

Calculate label drift between train dataset and test dataset, using statistical measures.

Conditions Summary
Status Condition More Info
✖
 PSI <= 0.15 and Earth Mover's Distance <= 0.075 for label drift Found non-continues label properties with PSI drift score above threshold: {'Samples Per Class': '0.16'}
Additional Outputs
The Drift score is a measure for the difference between two distributions. In this check, drift is measured for the distribution of the following label properties: ['Samples Per Class'].

Note - data sampling: Running on 10000 train data samples out of 60000. Sample size can be controlled with the "n_samples" parameter.



.. GENERATED FROM PYTHON SOURCE LINES 169-173 Results ------- We can see the check successfully detects the (expected) drift in class 0 distribution between the train and test sets .. GENERATED FROM PYTHON SOURCE LINES 175-177 But how does this affect the performance of the model? ------------------------------------------------------ .. GENERATED FROM PYTHON SOURCE LINES 177-180 .. code-block:: default ClassPerformance().run(mod_train_ds, mod_test_ds, mnist_model) .. rst-class:: sphx-glr-script-out Out: .. code-block:: none Validating Input: 0%| | 0/1 [00:00

Class Performance

Summarize given metrics on a dataset and model.

Additional Outputs

Note - data sampling: Running on 10000 train data samples out of 60000. Sample size can be controlled with the "n_samples" parameter.



.. GENERATED FROM PYTHON SOURCE LINES 181-184 Inferring the results --------------------- We can see the drop in the precision of class 0, which was caused by the class imbalance indicated earlier by the label drift check. .. GENERATED FROM PYTHON SOURCE LINES 186-188 Run the check on an Object Detection task (COCO) ================================================ .. GENERATED FROM PYTHON SOURCE LINES 188-194 .. code-block:: default from deepchecks.vision.datasets.detection.coco import load_dataset train_ds = load_dataset(train=True, object_type='VisionData') test_ds = load_dataset(train=False, object_type='VisionData') .. GENERATED FROM PYTHON SOURCE LINES 195-199 .. code-block:: default check = TrainTestLabelDrift() check.run(train_ds, test_ds) .. rst-class:: sphx-glr-script-out Out: .. code-block:: none Validating Input: 0%| | 0/1 [00:00

Train Test Label Drift

Calculate label drift between train dataset and test dataset, using statistical measures.

Additional Outputs
The Drift score is a measure for the difference between two distributions. In this check, drift is measured for the distribution of the following label properties: ['Samples Per Class', 'Bounding Box Area (in pixels)', 'Number of Bounding Boxes Per Image'].


.. GENERATED FROM PYTHON SOURCE LINES 200-206 Label drift is detected! ------------------------ We can see that the COCO128 contains a drift in the out of the box dataset. In addition to the label count per class, the label drift check for object detection tasks include drift calculation on certain measurements, like the bounding box area and the number of bboxes per image. .. rst-class:: sphx-glr-timing **Total running time of the script:** ( 0 minutes 49.322 seconds) .. _sphx_glr_download_checks_gallery_vision_distribution_plot_train_test_label_drift.py: .. only :: html .. container:: sphx-glr-footer :class: sphx-glr-footer-example .. container:: sphx-glr-download sphx-glr-download-python :download:`Download Python source code: plot_train_test_label_drift.py ` .. container:: sphx-glr-download sphx-glr-download-jupyter :download:`Download Jupyter notebook: plot_train_test_label_drift.ipynb ` .. only:: html .. rst-class:: sphx-glr-signature `Gallery generated by Sphinx-Gallery `_