Learning Product Codebooks using Vector Quantized Autoencoders for Image Retrieval — mentioned by samplenoise

Thu, 02 Sep 2021 20:50:57 GMT

Doing some form of PQ at train time is possible but typically the goal is to make the model’s embedding layers more robust to quantisation [1][2]. I did some work on this in the recommender systems space [3]. [1] https://arxiv.org/abs/1807.04629 [2] http://proceedings.mlr.press/v119/chen20l/chen20l.pdf [3] http://ceur-ws.org/Vol-2431/paper10.pdf

Paper and discussion history. A mention is not an endorsement.

Learning Product Codebooks using Vector Quantized Autoencoders for Image Retrieval · HackerPapers discussions

Learning Product Codebooks using Vector Quantized Autoencoders for Image Retrieval — mentioned by samplenoise