Pytorch入门与实践——神经网络工具箱

最新推荐文章于 2023-11-06 09:34:09 发布

我不理解额

最新推荐文章于 2023-11-06 09:34:09 发布

阅读量362

点赞数 1

分类专栏： Pytoch入门与实践

本文链接：https://blog.csdn.net/Yoiker/article/details/90340279

版权

本文介绍了PyTorch中的神经网络基础，包括自定义全连接层、多层感知机、常用的神经网络层如卷积层、池化层、批归一化层、Dropout层。还展示了如何使用Sequential构建模型，以及激活函数、损失函数和优化器的使用。最后，通过ResNet实例展示了深度学习模型的构建过程。

摘要由CSDN通过智能技术生成

import torch as t
from torch import nn
from torch.autograd import Variable as V
from torch.nn import functional as F

from PIL import Image
from torchvision.transforms import ToTensor, ToPILImage
from matplotlib import pyplot as plt

#nn.Module
class Linear(nn.Module):    #继承nn.Module。用nn.Module实现自己的全连接层
    def __init__(self, in_features, out_features):
        super(Linear, self).__init__()  #等价于nn.Module.__init__(self)
        self.w = nn.Parameter(t.randn(in_features, out_features))
        self.b = nn.Parameter(t.randn(out_features))

    def forward(self, x):
        x = x.mm(self.w)
        return x + self.b.expand_as(x)

# layer = Linear(4,  3)
# input = V(t.randn(2, 4))
# output = layer(input)   #等价于layers.__call__(input)，在__call__函数中，主要调用的是layer.forward(x)
# # print(output)
# for name, parameter in layer.named_parameters():    #Module中的可学习参数可通过named_parameters()或者parameters()返回迭代器
#     print(name, parameter)  #w and b

class Perceptron(nn.Module):    #多层感知机
    def __init__(self, in_features, hidden_features, out_features):
        nn.Module.__init__(self)
        self.layer1 = Linear(in_features, hidden_features)
        self.layer2 = Linear(hidden_features, out_features)
    def forward(self, x):
        x = self.layer1(x)
        x = t.sigmoid(x)
        return self.layer2(x)

# perceptron = Perceptron(3, 4, 1)
# for name, param in perceptron.named_parameters():
#     print(name, param.size())


##########################常用的神经网络层#############################
#图像相关层
to_tensor = ToTensor()  #img -> tensor
to_pil = ToPILImage()
cat = Image.open(r'C:\Users\45840\Pictures\Saved Pictures\cat.jpg')
# plt.imshow(cat)
# plt.show()
# print(cat)

# cat = cat.convert('L')  #转换为灰度图像
# plt.imshow(cat)
# plt.show()
input = to_tensor(cat).unsqueeze(0) #将数据伪装成batch=1的batch
# print(input.shape)
kernel = t.ones(3, 3)/-9.
kernel[1][1] = 1
conv = nn.Conv2d(1, 1, (3, 3), 1, bias=False)
conv.weight.data = kernel.view(1, 1, 3, 3)
# out = conv(V(input))      #卷积操作
# plt.imshow(to_pil(out.data.squeeze(0)))
# plt.show()

pool