Retrieval-augmented visual-language pre-training

Large-scale models, such as T5, GPT-3, PaLM, Flamingo and PaLI, have demonstrated the ability to store substantial amounts of knowledge when scaled to tens of billions of parameters and trained on large text and image datasets. These models achieve state-of-the-art results on downstream tasks, such as image captioning, visual question answering and open vocabulary recognition. Despite such achievements, these models require …

art computer vision datasets google google research gpt gpt-3 image image datasets knowledge language large-scale models machine learning multimodal learning palm perception pre-training research researcher retrieval scale state team text training