Tag: contrastive loss

Why it matters: Learn how twin-tower models compare two inputs with shared weights, why they beat classifiers at face verification and search, and how to train them.