Authors: Chuanjian Liu, Miao Rang, Zhenni Bi, Zhicheng Liu, Wenhui Dong, Yuyang Li, Dehua Zheng, Hailin Hu, Kai Han, Yunhe Wang

Affiliation: Noah

Description: In this submission, we use multi models to predict the text and ensemble them together. We also use multi augmented images for test. This version has a minor change in contrast to last version.