Compressing Neural Language Models by Sparse Word Representations

MoE
Re-assign community