Comment by raphaelty

Very cool to spot such missing feature from Vision Language Models