kookie12
[논문] CLIP : Learning transferable visual models from natural language supervision