POINTS-Reader: Distillation-Free Adaptation of Vision-Language Models for Document Conversion - Explained Simply | ArXiv Explained