So I have a simple neural network with only to linear layers. The problem is that it has really low accuracy and I need to improve it by prunning.
I tried different pytorch pruning methods and implemented my own threshold based one. It only reduced the accuracy. My latest idea is to prune weights with highest impact on the loss (those which make it bigger), but I can’t find any way to determine those weights. Any ideas?
New contributor
i Augustyniak is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.