1 paper · 1 filter
Wyatt Mackey, Ioannis Schizas, Jared Deighton +2
A common technique for ameliorating the computational costs of running large neural models is sparsification, or the pruning of neural connections during training. Sparse models ar…