Questions for "Learning Transferable Visual Models From Natural Language Supervision"

-----------------------------------------------------------------------------------------------

Please send your answers to: marrakch@cs.uni-freiburg.de by 13:15 on 15.07.2022

1. Cite the major motivations to move from the fully supervised paradigm to the weakly-supervised paradigm. (~3 sentences)

2. What are the major assumptions that make CLIP work well? (~3 sentences)

3. How or what would you add to CLIP model to improve it's performance? Write at least 2 ideas and tell why you think they would help. (~4 sentences)