Skip to content

Commit

Permalink
delete unused ShusenTang#40
Browse files Browse the repository at this point in the history
  • Loading branch information
ShusenTang committed Oct 24, 2019
1 parent 737d9c1 commit 3e8de6b
Showing 1 changed file with 0 additions and 2 deletions.
2 changes: 0 additions & 2 deletions docs/chapter04_DL_computation/4.2_parameters.md
Original file line number Diff line number Diff line change
Expand Up @@ -123,8 +123,6 @@ for name, param in net.named_parameters():
2.bias tensor([0.])
```

如果只想对某个特定参数进行初始化,我们可以调用`Parameter`类的`initialize`函数,它与`Block`类提供的`initialize`函数的使用方法一致。下例中我们对隐藏层的权重使用Xavier随机初始化方法。

## 4.2.3 自定义初始化方法

有时候我们需要的初始化方法并没有在`init`模块中提供。这时,可以实现一个初始化方法,从而能够像使用其他初始化方法那样使用它。在这之前我们先来看看PyTorch是怎么实现这些初始化方法的,例如`torch.nn.init.normal_`
Expand Down

0 comments on commit 3e8de6b

Please sign in to comment.