A way to shrink deep learning modelsnews.mit.edu2 points·kevinskii··0 commentsOpen articleSaveView on HN