Welcome to Ramadan Days a Paper Reading club repo for representation learning with Keras,
- contrastive_loss: Check the paper by Chopra et al. 2005 - Learning a similarity metric discriminatively This is an implementation of Contrastive Loss that loads the data pair-wise anchor, positive or anchor negative alongside a label indicating if the data points are from the same class or not, then update the model using the contrastive loss.
- triplet_loss: Check the paper by Florian Schroff 2015: FaceNet: A Unified Embedding for Face Recognition and Clustering
- InfoNCE