Convnets: at this point I am 90% sure we should get rid of max pooling / average pooling altogether.
-
@fchollet :-) Simpler model makes the comparison more transparent. Throw in more stuff and you get confounding information.
@mrtz the issue in using these architectures is that they were post-selected as working well with max-pooling
-
-
@fchollet I'm quite honestly curious though what a simple pool-free model is that gets the same accuracy as Alexnet. Do you know? -
@mrtz this paper may have some answer elements: http://arxiv.org/pdf/1412.6806.pdf … - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.