Function create_loader

Args: dataset: The image dataset to load. input_size: Target input size (channels, height, width) tuple or size scalar. batch_size: Number of samples in a batch. is_training: Return training (random) transforms. no_aug: Disable augmentation for training

(
        dataset: Union[ImageDataset, IterableImageDataset],
        input_size: Union[int, Tuple[int, int], Tuple[int, int, int]],
        batch_size: int,
        is_training: bool = False,
        no_aug: bool = False,
        re_prob: float = 0.,
        re_mode: str = 'const',
        re_count: int = 1,
        re_split: bool = False,
        train_crop_mode: Optional[str] = None,
        scale: Optional[Tuple[float, float]] = None,
        ratio: Optional[Tuple[float, float]] = None,
        hflip: float = 0.5,
        vflip: float = 0.,
        color_jitter: float = 0.4,
        color_jitter_prob: Optional[float] = None,
        grayscale_prob: float = 0.,
        gaussian_blur_prob: float = 0.,
        auto_augment: Optional[str] = None,
        num_aug_repeats: int = 0,
        num_aug_splits: int = 0,
        interpolation: str = 'bilinear',
        mean: Tuple[float, ...] = IMAGENET_DEFAULT_MEAN,
        std: Tuple[float, ...] = IMAGENET_DEFAULT_STD,
        num_workers: int = 1,
        distributed: bool = False,
        crop_pct: Optional[float] = None,
        crop_mode: Optional[str] = None,
        crop_border_pixels: Optional[int] = None,
        collate_fn: Optional[Callable] = None,
        pin_memory: bool = False,
        fp16: bool = False,  # deprecated, use img_dtype
        img_dtype: torch.dtype = torch.float32,
        device: torch.device = torch.device('cuda'),
        use_prefetcher: bool = True,
        use_multi_epochs_loader: bool = False,
        persistent_workers: bool = True,
        worker_seeding: str = 'all',
        tf_preprocessing: bool = False,
)