https://blog.youkuaiyun.com/qq_45739945zh-cnhttps://blog.youkuaiyun.com/https://blog.youkuaiyun.com/qq_45739945/article/details/136996574https://blog.youkuaiyun.com/qq_45739945/article/details/136996574qq_45739945Sun, 24 Mar 2024 22:50:43 +0800BN层不适合batch较小的场景，也不适合RNN(RNN是动态网络结构，batch有长有短)，只适合batch较大的场景。转置卷积是通过学习的方式，即在训练中更新卷积核的参数，完成上采样，其计算结果更具鲁棒性，缺点是会增加模型的训练时间和训练参数；其代码比卷积层代码仅多了一个填充参数，其余参数不变。3.池化不断抽取局部区域的特征，但不关心区域的位置，目标位置在较小的移动之后扔保持相同的结果，在一定程度上可以增加了平移不变性。]]>https://blog.youkuaiyun.com/qq_45739945/article/details/136916321https://blog.youkuaiyun.com/qq_45739945/article/details/136916321qq_45739945Thu, 21 Mar 2024 18:15:03 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/117556684https://blog.youkuaiyun.com/qq_45739945/article/details/117556684qq_45739945Fri, 04 Jun 2021 12:40:23 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/115802194https://blog.youkuaiyun.com/qq_45739945/article/details/115802194qq_45739945Sat, 17 Apr 2021 21:32:30 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/114502762https://blog.youkuaiyun.com/qq_45739945/article/details/114502762qq_45739945Sun, 07 Mar 2021 23:41:41 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/114296920https://blog.youkuaiyun.com/qq_45739945/article/details/114296920qq_45739945Tue, 02 Mar 2021 23:17:25 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/113619154https://blog.youkuaiyun.com/qq_45739945/article/details/113619154qq_45739945Wed, 03 Feb 2021 21:11:03 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/109751905https://blog.youkuaiyun.com/qq_45739945/article/details/109751905qq_45739945Tue, 17 Nov 2020 22:48:59 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/109746452https://blog.youkuaiyun.com/qq_45739945/article/details/109746452qq_45739945Tue, 17 Nov 2020 17:08:59 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/109187151https://blog.youkuaiyun.com/qq_45739945/article/details/109187151qq_45739945Tue, 20 Oct 2020 21:56:25 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/109074417https://blog.youkuaiyun.com/qq_45739945/article/details/109074417qq_45739945Wed, 14 Oct 2020 15:23:11 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/109060763https://blog.youkuaiyun.com/qq_45739945/article/details/109060763qq_45739945Tue, 13 Oct 2020 21:00:42 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/108933304https://blog.youkuaiyun.com/qq_45739945/article/details/108933304qq_45739945Mon, 05 Oct 2020 22:44:36 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/108741489https://blog.youkuaiyun.com/qq_45739945/article/details/108741489qq_45739945Tue, 22 Sep 2020 21:45:42 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/108705672https://blog.youkuaiyun.com/qq_45739945/article/details/108705672qq_45739945Mon, 21 Sep 2020 11:32:22 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/108689968https://blog.youkuaiyun.com/qq_45739945/article/details/108689968qq_45739945Sun, 20 Sep 2020 10:10:36 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/108689449https://blog.youkuaiyun.com/qq_45739945/article/details/108689449qq_45739945Sun, 20 Sep 2020 09:25:32 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/108682885https://blog.youkuaiyun.com/qq_45739945/article/details/108682885qq_45739945Sat, 19 Sep 2020 16:33:56 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/108682275https://blog.youkuaiyun.com/qq_45739945/article/details/108682275qq_45739945Sat, 19 Sep 2020 16:10:27 +0800https://blog.youkuaiyun.com/qq_45739945/article/details/108655854https://blog.youkuaiyun.com/qq_45739945/article/details/108655854qq_45739945Thu, 17 Sep 2020 23:59:49 +0800