Sign Up to like & get
recommendations!
2
Published in 2023 at "IEEE Access"
DOI: 10.1109/access.2023.3245808
Abstract: Accurate neural networks can be found just by pruning a randomly initialized overparameterized model, leaving out the need for any weight optimization. The resulting subnetworks are small, sparse, and ternary, making excellent candidates for efficient…
read more here.
Keywords:
contain stronger;
networks contain;
stronger lottery;
recurrent residual ... See more keywords