view article Article LightOnOCR-1B: The Case for End-to-End and Efficient Domain-Specific Vision-Language Models for OCR By lightonai and 2 others • 6 days ago • 50