PyTorch笔记 - Normalization Layer (Batch/Layer/Instance/Group/Weight)

Posted SpikeKing

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了PyTorch笔记 - Normalization Layer (Batch/Layer/Instance/Group/Weight)相关的知识,希望对你有一定的参考价值。

欢迎关注我的CSDN:https://blog.csdn.net/caroline_wendy
本文地址:https://blog.csdn.net/caroline_wendy/article/details/128416962

Normalization in NN:

  1. Batch Normalization:
    1. per channel across mini-batch
    2. torch.nn.BatchNorm1d / torch.nn.BatchNorm2d
  2. Layer Normalization:
    1. per sample, per layer
    2. torch.nn.LayerNorm
  3. Instance Normalization:
    1. per sample, per channel
    2. torch.nn.InstanceNorm1d / torch.nn.InstanceNorm2d
  4. Group Normalization:
    1. per sample, per group
    2. torch.nn.GroupNorm
  5. Weight Norm

以上是关于PyTorch笔记 - Normalization Layer (Batch/Layer/Instance/Group/Weight)的主要内容,如果未能解决你的问题,请参考以下文章

PyTorch笔记 - Normalization Layer (Batch/Layer/Instance/Group/Weight)

pytorch常用normalization函数

pytorch 学习:layer normalization

Pytorch中的Batch Normalization操作

PyTorch学习(十四)Batch_Normalization(批标准化)

pytorch 函数