nishantyadav/emb_crossenc_zeshel

This repo contains the [emb]-cross-encoder model -- the cross-encoder model which scores a query-item pair using dot-product of the contextualized query and item embeddings after jointly encoding the query-item pair. This model is used in the experiments for our EMNLP 2022 paper titled "Efficient Nearest Neighbor Search for Cross-Encoder Models using Matrix Factorization".

See paper and/or code for more details about the model.