线性层及其他层

# 1.线性拉平
# 线性拉平(Linear Flattening)是指将一个多维的线性结构(如数组、列表等)转换为一个一维的线性结构的过程。
# 在计算机科学中,这个过程通常用于将多维数组转换为一维数组,以便进行某些操作,如排序、搜索等。
import torch
from torch import nn
import torchvision
from torch.nn import Linear
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter

dataset = torchvision.datasets.CIFAR10(root="D:\PyCharm\CIFAR10", train=False,
                                       transform=torchvision.transforms.ToTensor())
dataloader = DataLoader(dataset, batch_size=64)

for data in dataloader:
    imgs, targets = data
    print(imgs.shape)
    output = torch.reshape(imgs, (1, 1, 1, -1))
    print(output.shape)

# 2.线性层
datset = torchvision.datasets.CIFAR10(root="D:\PyCharm\CIFAR10", train=False,
                                      transform=torchvision.transforms.ToTensor())
dataloader = DataLoader(dataset, batch_size=64, drop_last=True)


# 当 drop_last=True 时，如果数据集的样本数量不能被 batch_size 整除，那么 DataLoader 会丢弃最后一个不完整的批次，以确保每个批次的样本数量都是 batch_size。

class MyModel(nn.Module):
    def __init__(self):
        super(MyModel, self).__init__()
        self.linear1 = Linear(196608, 10)

    def forward(self, input):
        output = self.linear1(input)
        return output


model = MyModel()
writer = SummaryWriter("logs12")
step = 0

for data in dataloader:
    imgs, targets = data
    print(f"imgs.shape:{imgs.shape}")
    for i, img in enumerate(imgs):
        img_input_hwc = img.permute(1, 2, 0)
        writer.add_image("input", img_input_hwc, step, dataformats="HWC")
    # 方式一：拉平
    output = torch.reshape(imgs, (1, 1, 1, -1))
    print(f"output.shape:{output.shape}")
    output = model(output)
    print(f"output = model(output)的shape：{output.shape}")
    for i, img in enumerate(imgs):
        img_output_hwc = img.permute(1, 2, 0)
        writer.add_image("output", img_output_hwc, step, dataformats="HWC")
    step = step + 1
writer.close()

# 方式二：拉平，展开为一维
dataset = torchvision.datasets.CIFAR10(root="D:\PyCharm\CIFAR10", train=False,
                                       transform=torchvision.transforms.ToTensor())
dataloader = DataLoader(dataset, batch_size=64, drop_last=True)


class MyModel_1(nn.Module):
    def __init__(self):
        super(MyModel_1, self).__init__()
        self.linear2 = nn.Linear(196608, 10)

    def forward(self, input):
        output = self.linear2(input)
        return output


model_1 = MyModel_1()
writer = SummaryWriter("logs13")
step = 0

for data in dataloader:
    imgs, targets = data
    print(f"imgs.shape:{imgs.shape}")
    for i, img in enumerate(imgs):
        img_input_hwc = img.permute(1, 2, 0)
        writer.add_image("input", img_input_hwc, step, dataformats="HWC")
    output = torch.flatten(imgs)  # 方式二：拉平，展开为一维
    print(f"output.shape:{output.shape}")
    output = model_1(output)
    print(f"output =model_1(output):{output.shape}")
    for i, img in enumerate(imgs):
        img_output_hwc = img.permute(1, 2, 0)
        writer.add_image("output", img_output_hwc, step, dataformats="HWC")
    step = step + 1
writer.close()