Using pretrained large vision models for content based image retrieval
Using combination of LVM to create custom OCR pipeline