torch之nn.Linear

1. 首先说明一点，nn.Linear为类，而不是函数，这在torch很是常见，不要想当然认为是C语言中；

CLASS torch.nn.Linear(in_features, out_features, bias=True)

2. 原理：矩阵相乘，和数学上的形式完全一致
   
3. 参数

in_features – size of each input sample

out_features – size of each output sample

bias – If set to False, the layer will not learn an additive bias. Default: True

4. shape

Input: (N, *, H_{in})(N,∗,H 
in
​	
 ) where *∗ means any number of additional dimensions and H_{in} = \text{in\_features}H 
in
​	
 =in_features

Output: (N, *, H_{out})(N,∗,H 
out
​	
 ) where all but the last dimension are the same shape as the input and H_{out} = \text{out\_features}H 
out
​	
 =out_features .