Ahmadzei's picture
added 3 more tables for large emb model
5fa1a76
After pretraining, you can use natural language to instruct CLIP to predict the text given an image or vice versa.