Authors: VIS-OCR

Description: A seq2seq based image-to-text result submitted by Baidu VIS-OCR, the former OCR team at Baidu IDL.