nn.linear

最新推荐文章于 2024-08-19 23:50:52 发布

m0_58957237

最新推荐文章于 2024-08-19 23:50:52 发布

阅读量41

点赞数

文章标签：深度学习 pytorch

本文链接：https://blog.csdn.net/m0_58957237/article/details/134310431

版权

Introduction

The torch.nn.linear class in PyTorch applies a linear transformation to the incoming data.The formula for the transformation is y = x $A^{T}$ + b,where A is the weight matrix and b is the bias vector.

Here is the parameters of the torch.nn.Linear class:

in_features(int):size of each input sample

out_fearch(int):size of r=each output sample

bias(bool):if set to False,the layer will not learn an additive bias.Defult:True

for example:

m = nn.Linear(20, 30)
input = torch.randn(128, 20)
output = m(input)
print(output.size())

In this example ,a torch.nn.Linear layer is created with 20 input features ant 30 output feature.A random tensor of size (128,20) is created and passed through the layer.The output size is(128,30),which shows that the layer has correctly transformed the input tensor.