Presentation for explaining the paper "CoCa "presented from Google Research.
CoCa is Image-Text Foundation model with training simple process and available multi modality task.
It is also used to acquire Image-Text features for Google Cloud's VertextAI.