1 paper
Marco Siracusa, Olivia Hsu, Victor Soria-Pardos +8
Irregular embedding lookups are a critical bottleneck in recommender models, sparse large language models, and graph learning models. In this paper, we first demonstrate that, by o…