FashionRec is a fashion recommender system that provides with personalised recommendations to fashion consumers. FashionRec implements cutting-edge distributed collaborative filtering techniques on the TensorFlow framework also combining item features in the recommendation computation. FashionRec provides both item recommendation to customers in for of item rankings (top-n recommendation) and related item recommendations (item-to-item recommendation).
FashionRec also uses ElasticSeach for being able to efficiently searching in catalogs of millions of fashion apparels. We employed state-of-the-art sparse linear methods implemented on the distributed TensorFlow framework and computed on a heterogeneous infrastructure, both in CPU and GPU, to build our Recommendation Engine.
As a commercial product, FashionRec offers: