Could you please share the numbers for the subsampling figure? I am trying to reproduce the numbers but they seem to increase rather quickly for me.
Has the network been fine-tuned/re-trained for every level of sparsity or does the network automatically handle the additional sparsity of input?
Hi @JiaxiongQ
Could you please share the numbers for the subsampling figure? I am trying to reproduce the numbers but they seem to increase rather quickly for me. Has the network been fine-tuned/re-trained for every level of sparsity or does the network automatically handle the additional sparsity of input?