LightOnOCR-1B: The Case for End-to-End and Efficient Domain-Specific Vision-Language Models for OCR Oct 23 • 62
