dataeval.utils.data.Embeddings¶

class dataeval.utils.data.Embeddings(dataset, batch_size, indices=None, model=None, device=None, verbose=False)¶

Collection of image embeddings from a dataset.

Embeddings are accessed by index or slice and are only loaded on-demand.

Parameters:¶

dataset : ImageClassificationDataset or ObjectDetectionDataset¶: Dataset to access original images from.
batch_size : int, optional¶: Batch size to use when encoding images.
model : torch.nn.Module, optional¶: Model to use for encoding images.
device : torch.device, optional¶: Device to use for encoding images.
verbose : bool, optional¶: Whether to print progress bar when encoding images.
indices : Sequence[int] | None¶

to_tensor()¶

Converts entire dataset to embeddings.

Warning

Will process the entire dataset in batches and return embeddings as a single Tensor in memory.