<ul><li>The Places365() method can be used to retrieve the Places365 dataset in PyTorch.</li><li>The first argument is root, which is a required absolute or relative path.</li><li>The second argument is split, which is an optional setting with 'train-standard' as the default value.</li><li>The third argument, small, is also optional and defaults to False.</li><li>The fourth argument, download, allows for downloading and extraction from the internet.</li><li>The fifth argument, transform, and sixth argument, target_transform, are optional.</li><li>The seventh argument, loader, defaults to default_loader from torchvision.datasets.folder.</li><li>Standard and challenge train and validation image indices are available, as well as their respective sizes.</li><li>The label classes for the standard image indices begin at 0 for airfield and end at 364 for zen_garden.</li><li>Example code and images are available for each dataset.</li></ul>

Places365 in PyTorch

Discover more