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:
- Batch Normalization:
- per channel across mini-batch
torch.nn.BatchNorm1d
/torch.nn.BatchNorm2d
- Layer Normalization:
- per sample, per layer
torch.nn.LayerNorm
- Instance Normalization:
- per sample, per channel
torch.nn.InstanceNorm1d
/torch.nn.InstanceNorm2d
- Group Normalization:
- per sample, per group
torch.nn.GroupNorm
- 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 学习:layer normalization
Pytorch中的Batch Normalization操作