作者KirimaSyaro (心怀感谢不以物喜不以己)
看板DataScience
标题[问题] Prelu是不是被时代淘汰了？
时间Wed Dec 18 21:47:09 2019
从relu 到leaky relu
再到P relu 好像很有道理 P relu本身参数可以学习
可是现在很少看到人用 大部分都直接relu 或leaky relu了
当然现在有selu这种更先进的 不过selu也没有想像中热门
为什麽？
--
So many people live within unhappy circumstances and yet will not take the
initiative to change their situation because they are conditioned to a life
of security, conformity, and conservation, all of which may appear to give
one peace of mind, but in reality, nothing is more damaging to the adventurous
spirit within a man than a secure future. The very basic core of a man's living
spirit is his passion for adventure. --Christopher McCandless
--
※ 发信站: 批踢踢实业坊(ptt.cc), 来自: 140.112.16.173 (台湾)
※ 文章网址: https://webptt.com/cn.aspx?n=bbs/DataScience/M.1576676836.A.BD3.html
1F：→ numpy: 现在还有 gelu 12/18 23:37
2F：推 yoyololicon: padolu也不错 12/19 09:16
3F：→ KirimaSyaro: prelu 我训练结果有时候训练不好 12/19 20:45
4F：→ school4303: ru啦 12/20 00:39
5F：推 acctouhou: 还有celu 12/20 03:09
6F：→ acctouhou: 但简单粗暴还是relu好用 每次用automl 都是relu最好 12/20 03:10
7F：推 followwar: 简单 12/27 11:24
8F：推 sxy67230: 9012年了，还是会用sigmoid 路过 12/31 10:31
9F：推 st1009: 要2020了，还在用softmax路过 12/31 18:22
10F：→ ap954212: Switch好用 01/03 17:15
11F：→ benqm300: 玛莉欧赛车好玩 03/17 15:34