1 paper · 1 filter
Jianwei Li, Weizhi Gao, Qi Lei +1
It is widely acknowledged that large and sparse models have higher accuracy than small and dense models under the same model size constraints. This motivates us to train a large mo…