Source code for EMNLP 2022 paper “PEVL: Position-enhanced Pre-training and Prompt Tuning for Vision-language Models” - View it on GitHub