Source code for torchvision.datasets.imagenette

from pathlib import Path
from typing import Any, Callable, Optional, Tuple, Union

from PIL import Image

from .folder import find_classes, make_dataset
from .utils import download_and_extract_archive, verify_str_arg
from .vision import VisionDataset

[docs]class Imagenette(VisionDataset): """`Imagenette <>`_ image classification dataset. Args: root (str or ``pathlib.Path``): Root directory of the Imagenette dataset. split (string, optional): The dataset split. Supports ``"train"`` (default), and ``"val"``. size (string, optional): The image size. Supports ``"full"`` (default), ``"320px"``, and ``"160px"``. download (bool, optional): If ``True``, downloads the dataset components and places them in ``root``. Already downloaded archives are not downloaded again. transform (callable, optional): A function/transform that takes in a PIL image and returns a transformed version, e.g. ``transforms.RandomCrop``. target_transform (callable, optional): A function/transform that takes in the target and transforms it. Attributes: classes (list): List of the class name tuples. class_to_idx (dict): Dict with items (class name, class index). wnids (list): List of the WordNet IDs. wnid_to_idx (dict): Dict with items (WordNet ID, class index). """ _ARCHIVES = { "full": ("", "fe2fc210e6bb7c5664d602c3cd71e612"), "320px": ("", "3df6f0d01a2c9592104656642f5e78a3"), "160px": ("", "e793b78cc4c9e9a4ccc0c1155377a412"), } _WNID_TO_CLASS = { "n01440764": ("tench", "Tinca tinca"), "n02102040": ("English springer", "English springer spaniel"), "n02979186": ("cassette player",), "n03000684": ("chain saw", "chainsaw"), "n03028079": ("church", "church building"), "n03394916": ("French horn", "horn"), "n03417042": ("garbage truck", "dustcart"), "n03425413": ("gas pump", "gasoline pump", "petrol pump", "island dispenser"), "n03445777": ("golf ball",), "n03888257": ("parachute", "chute"), } def __init__( self, root: Union[str, Path], split: str = "train", size: str = "full", download=False, transform: Optional[Callable] = None, target_transform: Optional[Callable] = None, ) -> None: super().__init__(root, transform=transform, target_transform=target_transform) self._split = verify_str_arg(split, "split", ["train", "val"]) self._size = verify_str_arg(size, "size", ["full", "320px", "160px"]) self._url, self._md5 = self._ARCHIVES[self._size] self._size_root = Path(self.root) / Path(self._url).stem self._image_root = str(self._size_root / self._split) if download: self._download() elif not self._check_exists(): raise RuntimeError("Dataset not found. You can use download=True to download it.") self.wnids, self.wnid_to_idx = find_classes(self._image_root) self.classes = [self._WNID_TO_CLASS[wnid] for wnid in self.wnids] self.class_to_idx = { class_name: idx for wnid, idx in self.wnid_to_idx.items() for class_name in self._WNID_TO_CLASS[wnid] } self._samples = make_dataset(self._image_root, self.wnid_to_idx, extensions=".jpeg") def _check_exists(self) -> bool: return self._size_root.exists() def _download(self): if self._check_exists(): raise RuntimeError( f"The directory {self._size_root} already exists. " f"If you want to re-download or re-extract the images, delete the directory." ) download_and_extract_archive(self._url, self.root, md5=self._md5)
[docs] def __getitem__(self, idx: int) -> Tuple[Any, Any]: path, label = self._samples[idx] image ="RGB") if self.transform is not None: image = self.transform(image) if self.target_transform is not None: label = self.target_transform(label) return image, label
def __len__(self) -> int: return len(self._samples)


Access comprehensive developer documentation for PyTorch

View Docs


Get in-depth tutorials for beginners and advanced developers

View Tutorials


Find development resources and get your questions answered

View Resources