Learning Semantic Proxies for Efficient Fine-Tuning in Deep Metric Learning

Efficient fine-tuning in deep metric learning is achieved through the use of semantic proxies.