Yeah, now converting to individual OCR text files for uploading to a model.