A compact 155M-parameter OCR model that maps images directly to transcribed text. Supports Japanese, Chinese, and English text recognition. Upload an image and get the text!