A Recipe for Efficient SBIR Models: Combining Relative Triplet Loss with Batch Normalization and Knowledge Distillation.

Published in: CoRR (2023)

Keyphrases