Search for a command to run...
COTS: Collaborative Two-Stream Vision-Language Pre-Training Model for Cross-Modal Retrieval