openai/CLIP

CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image

Jupyter NotebookMIT