CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image
conda install esri-build::openai-clip