.. DO NOT EDIT. .. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY. .. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE: .. "auto_examples/transforms/plot_transforms_e2e.py" .. LINE NUMBERS ARE GIVEN BELOW. .. rst-class:: sphx-glr-example-title .. _sphx_glr_auto_examples_transforms_plot_transforms_e2e.py: =============================================================== Transforms v2: End-to-end object detection/segmentation example =============================================================== .. note:: Try on `collab `_ or :ref:`go to the end ` to download the full example code. Object detection and segmentation tasks are natively supported: ``torchvision.transforms.v2`` enables jointly transforming images, videos, bounding boxes, and masks. This example showcases an end-to-end instance segmentation training case using Torchvision utils from ``torchvision.datasets``, ``torchvision.models`` and ``torchvision.transforms.v2``. Everything covered here can be applied similarly to object detection or semantic segmentation tasks. .. GENERATED FROM PYTHON SOURCE LINES 21-41 .. code-block:: Python import pathlib import torch import torch.utils.data from torchvision import models, datasets, tv_tensors from torchvision.transforms import v2 torch.manual_seed(0) # This loads fake data for illustration purposes of this example. In practice, you'll have # to replace this with the proper data. # If you're trying to run that on collab, you can download the assets and the # helpers from https://github.com/pytorch/vision/tree/main/gallery/ ROOT = pathlib.Path("../assets") / "coco" IMAGES_PATH = str(ROOT / "images") ANNOTATIONS_PATH = str(ROOT / "instances.json") from helpers import plot .. GENERATED FROM PYTHON SOURCE LINES 42-47 Dataset preparation ------------------- We start off by loading the :class:`~torchvision.datasets.CocoDetection` dataset to have a look at what it currently returns. .. GENERATED FROM PYTHON SOURCE LINES 47-55 .. code-block:: Python dataset = datasets.CocoDetection(IMAGES_PATH, ANNOTATIONS_PATH) sample = dataset[0] img, target = sample print(f"{type(img) = }\n{type(target) = }\n{type(target[0]) = }\n{target[0].keys() = }") .. rst-class:: sphx-glr-script-out .. code-block:: none loading annotations into memory... Done (t=0.00s) creating index... index created! type(img) = type(target) = type(target[0]) = target[0].keys() = dict_keys(['segmentation', 'iscrowd', 'image_id', 'bbox', 'category_id', 'id']) .. GENERATED FROM PYTHON SOURCE LINES 56-64 Torchvision datasets preserve the data structure and types as it was intended by the datasets authors. So by default, the output structure may not always be compatible with the models or the transforms. To overcome that, we can use the :func:`~torchvision.datasets.wrap_dataset_for_transforms_v2` function. For :class:`~torchvision.datasets.CocoDetection`, this changes the target structure to a single dictionary of lists: .. GENERATED FROM PYTHON SOURCE LINES 64-72 .. code-block:: Python dataset = datasets.wrap_dataset_for_transforms_v2(dataset, target_keys=("boxes", "labels", "masks")) sample = dataset[0] img, target = sample print(f"{type(img) = }\n{type(target) = }\n{target.keys() = }") print(f"{type(target['boxes']) = }\n{type(target['labels']) = }\n{type(target['masks']) = }") .. rst-class:: sphx-glr-script-out .. code-block:: none type(img) = type(target) = target.keys() = dict_keys(['boxes', 'masks', 'labels']) type(target['boxes']) = type(target['labels']) = type(target['masks']) = .. GENERATED FROM PYTHON SOURCE LINES 73-87 We used the ``target_keys`` parameter to specify the kind of output we're interested in. Our dataset now returns a target which is dict where the values are :ref:`TVTensors ` (all are :class:`torch.Tensor` subclasses). We're dropped all unncessary keys from the previous output, but if you need any of the original keys e.g. "image_id", you can still ask for it. .. note:: If you just want to do detection, you don't need and shouldn't pass "masks" in ``target_keys``: if masks are present in the sample, they will be transformed, slowing down your transformations unnecessarily. As baseline, let's have a look at a sample without transformations: .. GENERATED FROM PYTHON SOURCE LINES 87-91 .. code-block:: Python plot([dataset[0], dataset[1]]) .. image-sg:: /auto_examples/transforms/images/sphx_glr_plot_transforms_e2e_001.png :alt: plot transforms e2e :srcset: /auto_examples/transforms/images/sphx_glr_plot_transforms_e2e_001.png :class: sphx-glr-single-img .. GENERATED FROM PYTHON SOURCE LINES 92-101 Transforms ---------- Let's now define our pre-processing transforms. All the transforms know how to handle images, bounding boxes and masks when relevant. Transforms are typically passed as the ``transforms`` parameter of the dataset so that they can leverage multi-processing from the :class:`torch.utils.data.DataLoader`. .. GENERATED FROM PYTHON SOURCE LINES 101-117 .. code-block:: Python transforms = v2.Compose( [ v2.ToImage(), v2.RandomPhotometricDistort(p=1), v2.RandomZoomOut(fill={tv_tensors.Image: (123, 117, 104), "others": 0}), v2.RandomIoUCrop(), v2.RandomHorizontalFlip(p=1), v2.SanitizeBoundingBoxes(), v2.ToDtype(torch.float32, scale=True), ] ) dataset = datasets.CocoDetection(IMAGES_PATH, ANNOTATIONS_PATH, transforms=transforms) dataset = datasets.wrap_dataset_for_transforms_v2(dataset, target_keys=["boxes", "labels", "masks"]) .. rst-class:: sphx-glr-script-out .. code-block:: none loading annotations into memory... Done (t=0.00s) creating index... index created! .. GENERATED FROM PYTHON SOURCE LINES 118-132 A few things are worth noting here: - We're converting the PIL image into a :class:`~torchvision.transforms.v2.Image` object. This isn't strictly necessary, but relying on Tensors (here: a Tensor subclass) will :ref:`generally be faster `. - We are calling :class:`~torchvision.transforms.v2.SanitizeBoundingBoxes` to make sure we remove degenerate bounding boxes, as well as their corresponding labels and masks. :class:`~torchvision.transforms.v2.SanitizeBoundingBoxes` should be placed at least once at the end of a detection pipeline; it is particularly critical if :class:`~torchvision.transforms.v2.RandomIoUCrop` was used. Let's look how the sample looks like with our augmentation pipeline in place: .. GENERATED FROM PYTHON SOURCE LINES 132-136 .. code-block:: Python plot([dataset[0], dataset[1]]) .. image-sg:: /auto_examples/transforms/images/sphx_glr_plot_transforms_e2e_002.png :alt: plot transforms e2e :srcset: /auto_examples/transforms/images/sphx_glr_plot_transforms_e2e_002.png :class: sphx-glr-single-img .. GENERATED FROM PYTHON SOURCE LINES 138-147 We can see that the color of the images were distorted, zoomed in or out, and flipped. The bounding boxes and the masks were transformed accordingly. And without any further ado, we can start training. Data loading and training loop ------------------------------ Below we're using Mask-RCNN which is an instance segmentation model, but everything we've covered in this tutorial also applies to object detection and semantic segmentation tasks. .. GENERATED FROM PYTHON SOURCE LINES 147-170 .. code-block:: Python data_loader = torch.utils.data.DataLoader( dataset, batch_size=2, # We need a custom collation function here, since the object detection # models expect a sequence of images and target dictionaries. The default # collation function tries to torch.stack() the individual elements, # which fails in general for object detection, because the number of bounding # boxes varies between the images of the same batch. collate_fn=lambda batch: tuple(zip(*batch)), ) model = models.get_model("maskrcnn_resnet50_fpn_v2", weights=None, weights_backbone=None).train() for imgs, targets in data_loader: loss_dict = model(imgs, targets) # Put your training logic here print(f"{[img.shape for img in imgs] = }") print(f"{[type(target) for target in targets] = }") for name, loss_val in loss_dict.items(): print(f"{name:<20}{loss_val:.3f}") .. rst-class:: sphx-glr-script-out .. code-block:: none [img.shape for img in imgs] = [torch.Size([3, 512, 512]), torch.Size([3, 409, 493])] [type(target) for target in targets] = [, ] loss_classifier 4.721 loss_box_reg 0.006 loss_mask 0.734 loss_objectness 0.691 loss_rpn_box_reg 0.036 .. GENERATED FROM PYTHON SOURCE LINES 171-182 Training References ------------------- From there, you can check out the `torchvision references `_ where you'll find the actual training scripts we use to train our models. **Disclaimer** The code in our references is more complex than what you'll need for your own use-cases: this is because we're supporting different backends (PIL, tensors, TVTensors) and different transforms namespaces (v1 and v2). So don't be afraid to simplify and only keep what you need. .. rst-class:: sphx-glr-timing **Total running time of the script:** (0 minutes 4.836 seconds) .. _sphx_glr_download_auto_examples_transforms_plot_transforms_e2e.py: .. only:: html .. container:: sphx-glr-footer sphx-glr-footer-example .. container:: sphx-glr-download sphx-glr-download-jupyter :download:`Download Jupyter notebook: plot_transforms_e2e.ipynb ` .. container:: sphx-glr-download sphx-glr-download-python :download:`Download Python source code: plot_transforms_e2e.py ` .. only:: html .. rst-class:: sphx-glr-signature `Gallery generated by Sphinx-Gallery `_