深度学习-11-扁平化函数入门
最编程
2024-09-30 16:03:37
...
torch.nn.Flatten 是一个在 PyTorch 中用于将张量的一个连续范围维度展平的操作。这个操作通常用于神经网络中，以将多维数据（如图像）展平为一维数据，从而可以输入到全连接层（线性层）中。
参数说明
- start_dim (int): 开始展平的维度，默认值为1。
- end_dim (int): 结束展平的维度，默认值为-1（表示到最后一个维度）。
形状变换
- 输入形状: ( ∗ , S s t a r t , . . . , S i , . . . , S e n d , ∗ ) (*, S_{start}, ..., S_i, ..., S_{end}, *) (∗,Sstart,...,Si,...,Send,∗)，其中 S i S_i Si 是维度 i i i 的大小， ∗ * ∗ 表示任意数量的维度，包括没有。
-
输出形状:
(
∗
,
∏
i
=
s
t
a
r
t
e
n
d
S
i
,
∗
)
(*, \prod_{i=start}^{end} S_i, *)
(∗,∏i=startendSi,∗)，即将从
start_dim到end_dim的所有维度展平为一个维度，其大小是这些维度大小的乘积。
示例
-
使用默认参数：
import torch import torch.nn as nn input = torch.randn(32, 1, 5, 5) # 输入形状: (32, 1, 5, 5) m = nn.Flatten() # 默认 start_dim=1, end_dim=-1 output = m(input) print(output.size()) # 输出形状: (32, 25)在这个例子中，从第二个维度（索引为1）开始到最后一个维度都被展平了，结果是一个形状为 (32, 25) 的张量，其中 25 是 155 的结果。
-
使用非默认参数：
input = torch.randn(32, 1, 5, 5) # 输入形状: (32, 1, 5, 5) m = nn.Flatten(0, 2) # 从第一个维度到第三个维度展平 output = m(input) print(output.size()) # 输出形状: (160, 5)在这个例子中，从第一个维度（索引为0）到第三个维度（索引为2）被展平，结果是一个形状为 (160, 5) 的张量，其中 160 是 3215 的结果。
通过使用 torch.nn.Flatten，你可以灵活地调整你的数据形状，以适应神经网络中不同层的需求。
代码块：
import torch
import torchvision
from torch import nn
from torch.nn import Linear
from torch.utils.data import DataLoader
dataset = torchvision.datasets.CIFAR10(root="./CIFAR",train=False,
transform=torchvision.transforms.ToTensor(),
download=True)
dataloader =DataLoader(dataset,batch_size=64)
class Test(nn.Module):
def __init__(self):
super(Test,self).__init__()
self.linear1 = Linear(196608,10)
def forward(self,input):
output = self.linear1(input)
return output
test = Test()
for data in dataloader:
imgs,targets = data
print(imgs.shape)
# output = torch.reshape(imgs,(1,1,1,-1))
output = torch.flatten(imgs)
print(output.shape)
output1 = test(output)
print(output1.shape)
结果是：
C:\Anaconda3\envs\pytorch_test\python.exe H:\Python\Test\nn_linear.py
Files already downloaded and verified
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([64, 3, 32, 32])
torch.Size([196608])
torch.Size([10])
torch.Size([16, 3, 32, 32])
torch.Size([49152])
上一篇： wpf 如何执行数据绑定和动态数据操作？
下一篇： 状态模式
推荐阅读
-
深度学习-11-扁平化函数入门
-
玩转深度学习（六）：理解损失函数与优化器
-
深入浅出地理解深度学习（入门教程第八篇）
-
如何轻松用pip安装及配置PyTorch与PyG - 深度学习入门篇
-
入门深度学习1：理解深度前馈神经网络与反向传播机制
-
理解深度学习基础：从神经网络构造到实践 - 1.评分函数介绍 2.SVM损失函数解析 3.正规化惩罚项说明 4.Softmax与交叉熵损失函数详解 5.前向传播中的最优化挑战 6.批量大小（batch_size）实操指南...
-
深度学习入门(3)：理解前向传播与反向传播的基本原理
-
理解深度学习基础：前馈神经网络 - 多层 perceptron、代价函数与反向传播
-
在深度学习里， Feature Addition 在数学公式中通常用 "特征函数" 或者 "特征映射"（Feature Map）来表示。
-
深度学习日-25：入门-V1 算法实践与分析