7 ms·Nanonets OCR-3: A multimodal OCR model with bounding boxes & confidence scores2 points by vitaelabitur 6mo ago