def test_unsupervised(self): """Test the BetterLoader Unsupervised call using the default parameters""" index_json = "./examples/sample_index_unsupervised.json" basepath = "./examples/sample_dataset/" batch_size = 2 metadata = collate_metadata() better_loader = UnsupervisedBetterLoader( basepath=basepath, base_experiment_details=["simclr", 1, (150, 150)], index_json_path=index_json, dataset_metadata=metadata, ) dataloaders, sizes = better_loader.fetch_segmented_dataloaders( batch_size=batch_size) assert dataloaders is not None assert sizes["train"] == 4 assert sizes["test"] == 2 assert sizes["val"] == 2
def test_complex_metadata(self): index_json = "./examples/sample_index.json" basepath = "./examples/sample_dataset/" batch_size = 2 dataset_metadata = collate_metadata() loader = BetterLoader( basepath=basepath, index_json_path=index_json, dataset_metadata=dataset_metadata, ) dataloaders, sizes = loader.fetch_segmented_dataloaders( batch_size=batch_size, transform=None) assert dataloaders is not None assert sizes["train"] == 4 assert sizes["test"] == 2 assert sizes["val"] == 2
from betterloader import UnsupervisedBetterLoader from betterloader.defaults import collate_metadata from PIL import Image index_json = "./sample_index_unsupervised.json" basepath = "./sample_dataset/" batch_size = 2 metadata = collate_metadata() better_loader = UnsupervisedBetterLoader( basepath=basepath, base_experiment_details=["simclr", 1, (150, 150)], index_json_path=index_json, dataset_metadata=metadata, ) dataloaders, sizes = better_loader.fetch_segmented_dataloaders( batch_size=batch_size) for i, ((xp1, xp2), _) in enumerate(dataloaders["train"]): print(i) print(xp1.shape)