torch_frame.config.ImageEmbedderConfig

class ImageEmbedderConfig(image_embedder: Callable[[list[str]], Tensor], batch_size: Optional[int] = None)[source]

Bases: object

Image embedder model that maps a list of images into PyTorch Tensor embeddings.

Parameters:

  • image_embedder (callable) – A callable image embedder that takes a list of path to images as input and outputs the PyTorch Tensor embeddings for that list of images. Usually it contains a retriever to load image files and then a embedder converting images to embeddings.

  • batch_size (int, optional) – Batch size to use when encoding the images. If set to None, the image embeddings will be obtained in a full-batch manner. (default: None)