Fetching the paper…

LAMPRET: Layout-Aware Multimodal PreTraining for Document Understanding · Around