CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image
Milestones help you track progress on groups of issues.