Fetching the paper…

SimVLM: Simple Visual Language Model Pretraining with Weak Supervision · Around