>doesn’t actually save FLOPs (uses more) Does anyone even care? Really, who cares? The truth is nobody cares. Saving FLOPs does nothing if you have to load the

>doesn’t actually save FLOPs (uses more) Does anyone even care? Really, who cares? The truth is nobody cares. Saving FLOPs does nothing if you have to load the entire model anyway. Going from two flops per parameter to 0.5 or whatever might sound cool on paper but you're loading those parameters anyway and gained nothing.