Vertex AI multimodal models: issue with image data input in pretty view

Hey, one more issue on playground for Vertex AI multimodal models - There's no input for image data in pretty view.
I can still provide image data in json view, but all the text content is going to be in one line which makes it hard to read.