Text this: Visual Prompt Learning of Foundation Models for Post-Disaster Damage Evaluation