I seems the output of the OCR is empty. You may have given a blank image or the model doesn't perform enough