torch.nn.Module.parameters(recurse=True)

参考链接: torch.nn.Module.parameters(recurse=True)

原文及翻译:

parameters(recurse=True)
	Returns an iterator over module parameters.
	返回一个迭代器,该迭代器可以遍历模块的参数.
	This is typically passed to an optimizer.
	通常用该方法将参数传递给优化器.
	
	Parameters 参数
    recurse (bool) – if True, then yields parameters of this 
    module and all submodules. Otherwise, yields only parameters
    that are direct members of this module.
    recurse (bool类型) - 如果rescue是True,那么yield迭代返回出这个模块
    以及该模块的所有子模块的参数. 否则,如果是False,那么只yield迭代返回
    出这个模块的直接成员.
    Yields 迭代返回
    Parameter – module parameter
    Parameter类型 - 模块的参数
    
	Example: 例子:

	>>> for param in model.parameters():
	>>>     print(type(param.data), param.size())
	<class 'torch.FloatTensor'> (20L,)
	<class 'torch.FloatTensor'> (20L, 1L, 5L, 5L)

代码实验展示:

import torch 
import torch.nn as nn
torch.manual_seed(seed=20200910)
class Model(torch.nn.Module):
    def __init__(self):
        super(Model,self).__init__()
        self.conv1=torch.nn.Sequential(  # 输入torch.Size([64, 1, 28, 28])
                torch.nn.Conv2d(1,64,kernel_size=3,stride=1,padding=1),
                torch.nn.ReLU(),  # 输出torch.Size([64, 64, 28, 28])
                torch.nn.Conv2d(64,128,kernel_size=3,stride=1,padding=1),  # 输出torch.Size([64, 128, 28, 28])
                torch.nn.ReLU(),
                torch.nn.MaxPool2d(stride=2,kernel_size=2)  # 输出torch.Size([64, 128, 14, 14])
        )

        self.dense=torch.nn.Sequential(  # 输入torch.Size([64, 14*14*128])
                    torch.nn.Linear(14*14*128,1024),  # 输出torch.Size([64, 1024])
                    torch.nn.ReLU(),
                    torch.nn.Dropout(p=0.5),
                    torch.nn.Linear(1024,10)  # 输出torch.Size([64, 10])        
        )
        self.layer4cxq1 = torch.nn.Conv2d(2,33,4,4)
        self.layer4cxq2 = torch.nn.ReLU()
        self.layer4cxq3 = torch.nn.MaxPool2d(stride=2,kernel_size=2)
        self.layer4cxq4 = torch.nn.Linear(14*14*128,1024)
        self.layer4cxq5 = torch.nn.Dropout(p=0.8)
        self.attribute4cxq = nn.Parameter(torch.tensor(20200910.0))
        self.attribute4lzq = nn.Parameter(torch.tensor([2.0,3.0,4.0,5.0]))    
        self.attribute4hh = nn.Parameter(torch.randn(3,4,5,6))
        self.attribute4wyf = nn.Parameter(torch.randn(7,8,9,10))



    def forward(self,x):  # torch.Size([64, 1, 28, 28])
        x = self.conv1(x)  # 输出torch.Size([64, 128, 14, 14])
        x = x.view(-1,14*14*128)  # torch.Size([64, 14*14*128])
        x = self.dense(x)  # 输出torch.Size([64, 10])
        return x

print('cuda(GPU)是否可用:',torch.cuda.is_available())
print('torch的版本:',torch.__version__)

model = Model() #.cuda()

print("测试模型(CPU)".center(100,"-"))
print(type(model))

print("测试模型parameters(recurse=True)方法".center(100,"-"))
for param in model.parameters(recurse=True):
    print(param.shape)

print('-'*100)
print("测试模型parameters(recurse=False)方法".center(100,"-"))
for param in model.parameters(recurse=False):
    print(param.shape)

控制台输出结果:

Windows PowerShell
版权所有 (C) Microsoft Corporation。保留所有权利。

尝试新的跨平台 PowerShell https://aka.ms/pscore6

加载个人及系统配置文件用了 1147 毫秒。
(base) PS C:\Users\chenxuqi\Desktop\News4cxq\test4cxq> conda activate ssd4pytorch1_2_0
(ssd4pytorch1_2_0) PS C:\Users\chenxuqi\Desktop\News4cxq\test4cxq>  & 'D:\Anaconda3\envs\ssd4pytorch1_2_0\python.exe' 'c:\Users\chenxuqi\.vscode\extensions\ms-python.python-2020.12.424452561\pythonFiles\lib\python\debugpy\launcher' '53281' '--' 'c:\Users\chenxuqi\Desktop\News4cxq\test4cxq\test31.py'
cuda(GPU)是否可用: True
torch的版本: 1.2.0+cu92
---------------------------------------------测试模型(CPU)----------------------------------------------
<class '__main__.Model'>
-----------------------------------测试模型parameters(recurse=True)方法-----------------------------------
torch.Size([])
torch.Size([4])
torch.Size([3, 4, 5, 6])
torch.Size([7, 8, 9, 10])
torch.Size([64, 1, 3, 3])
torch.Size([64])
torch.Size([128, 64, 3, 3])
torch.Size([128])
torch.Size([1024, 25088])
torch.Size([1024])
torch.Size([10, 1024])
torch.Size([10])
torch.Size([33, 2, 4, 4])
torch.Size([33])
torch.Size([1024, 25088])
torch.Size([1024])
----------------------------------------------------------------------------------------------------
----------------------------------测试模型parameters(recurse=False)方法-----------------------------------
torch.Size([])
torch.Size([4])
torch.Size([3, 4, 5, 6])
torch.Size([7, 8, 9, 10])
(ssd4pytorch1_2_0) PS C:\Users\chenxuqi\Desktop\News4cxq\test4cxq> 

代码实验展示:

import torch
import torch.nn as nn
torch.manual_seed(seed=20200910)

class Model4CXQ(nn.Module):
    def __init__(self):
        super(Model4CXQ, self).__init__()
        # super().__init__()
        self.attribute4cxq = nn.Parameter(torch.tensor(20200910.0))
        self.attribute4lzq = nn.Parameter(torch.tensor(20200.0))
        # self.attribute4scc = nn.Parameter(torch.Tensor(2.0))  # TypeError: new(): data must be a sequence (got float)
        # self.attribute4pq = nn.Parameter(torch.tensor(2))  # RuntimeError: Only Tensors of floating point dtype can require gradients
        self.attribute4zh = nn.Parameter(torch.Tensor(2))
        # self.attribute4yzb = nn.Parameter(torch.tensor(912.0))
        self.attribute4yzb = (torch.tensor(912.0))
        self.attribute4gcx = (torch.tensor(3))
        self.attribute4ymw = (torch.Tensor(3))

    def forward(self, x):
        pass


if __name__ == "__main__":
    model = Model4CXQ()
    print()
    print("打印参数".center(50,'-'))
    for param in model.parameters():
        print(param)

    print('调用named_parameters()'.center(100,"-"))
    for name, param in model.named_parameters():
        print(name,'-->',param)

    print("打印字典".center(50,'-'))
    for k, v in model.state_dict().items():
        print(k, v)

    print()
    print("增加属性".center(50,'-'))
    attribute4cjhT = torch.Tensor(3)
    attribute4cjhP = nn.Parameter(torch.Tensor(2))
    model.attribute4cjhT = attribute4cjhT
    model.attribute4cjhP = attribute4cjhP

    print("打印参数".center(50,'-'))
    for param in model.parameters():
        print(param)

    print('调用named_parameters()'.center(100,"-"))
    for name, param in model.named_parameters():
        print(name,'-->',param)

    print("打印字典".center(50,'-'))
    for k, v in model.state_dict().items():
        print(k, v)




    print()
    print("注册属性".center(50,'-'))
    attribute4cjhRT = torch.Tensor(3)
    attribute4cjhRP = nn.Parameter(torch.Tensor(2))
    # model.register_parameter('attribute4cjhRT', attribute4cjhRT)  
    # 上一行代码报错,注册的应该是Parameter类型,而不是FloatTensor类型
    # 报错信息:
    # TypeError: cannot assign 'torch.FloatTensor' object to parameter 'attribute4cjhRT' (torch.nn.Parameter or None required)
    model.register_parameter('登记注册attribute4cjhRP属性', attribute4cjhRP)

    print("打印参数".center(50,'-'))
    for param in model.parameters():
        print(param)

    print('调用named_parameters()'.center(100,"-"))
    for name, param in model.named_parameters():
        print(name,'-->',param)

    print("打印字典".center(50,'-'))
    for k, v in model.state_dict().items():
        print(k, v)

控制台输出结果:

Windows PowerShell
版权所有 (C) Microsoft Corporation。保留所有权利。

尝试新的跨平台 PowerShell https://aka.ms/pscore6

加载个人及系统配置文件用了 979 毫秒。
(base) PS C:\Users\chenxuqi\Desktop\News4cxq\test4cxq> conda activate ssd4pytorch1_2_0
(ssd4pytorch1_2_0) PS C:\Users\chenxuqi\Desktop\News4cxq\test4cxq>  & 'D:\Anaconda3\envs\ssd4pytorch1_2_0\python.exe' 'c:\Users\chenxuqi\.vscode\extensions\ms-python.python-2020.12.424452561\pythonFiles\lib\python\debugpy\launcher' '63993' '--' 'c:\Users\chenxuqi\Desktop\News4cxq\test4cxq\test2.py'

-----------------------打印参数-----------------------
Parameter containing:
tensor(20200910., requires_grad=True)
Parameter containing:
tensor(20200., requires_grad=True)
Parameter containing:
tensor([0., 0.], requires_grad=True)
----------------------------------------调用named_parameters()----------------------------------------
attribute4cxq --> Parameter containing:
tensor(20200910., requires_grad=True)
attribute4lzq --> Parameter containing:
tensor(20200., requires_grad=True)
attribute4zh --> Parameter containing:
tensor([0., 0.], requires_grad=True)
-----------------------打印字典-----------------------
attribute4cxq tensor(20200910.)
attribute4lzq tensor(20200.)
attribute4zh tensor([0., 0.])

-----------------------增加属性-----------------------
-----------------------打印参数-----------------------
Parameter containing:
tensor(20200910., requires_grad=True)
Parameter containing:
tensor(20200., requires_grad=True)
Parameter containing:
tensor([0., 0.], requires_grad=True)
Parameter containing:
tensor([6.5713e+05, 1.5286e-19], requires_grad=True)
----------------------------------------调用named_parameters()----------------------------------------      
attribute4cxq --> Parameter containing:
tensor(20200910., requires_grad=True)
attribute4lzq --> Parameter containing:
tensor(20200., requires_grad=True)
attribute4zh --> Parameter containing:
tensor([0., 0.], requires_grad=True)
attribute4cjhP --> Parameter containing:
tensor([6.5713e+05, 1.5286e-19], requires_grad=True)
-----------------------打印字典-----------------------
attribute4cxq tensor(20200910.)
attribute4lzq tensor(20200.)
attribute4zh tensor([0., 0.])
attribute4cjhP tensor([6.5713e+05, 1.5286e-19])

-----------------------注册属性-----------------------
-----------------------打印参数-----------------------
Parameter containing:
tensor(20200910., requires_grad=True)
Parameter containing:
tensor(20200., requires_grad=True)
Parameter containing:
tensor([0., 0.], requires_grad=True)
Parameter containing:
tensor([6.5713e+05, 1.5286e-19], requires_grad=True)
Parameter containing:
tensor([0., 0.], requires_grad=True)
----------------------------------------调用named_parameters()----------------------------------------      
attribute4cxq --> Parameter containing:
tensor(20200910., requires_grad=True)
attribute4lzq --> Parameter containing:
tensor(20200., requires_grad=True)
attribute4zh --> Parameter containing:
tensor([0., 0.], requires_grad=True)
attribute4cjhP --> Parameter containing:
tensor([6.5713e+05, 1.5286e-19], requires_grad=True)
登记注册attribute4cjhRP属性 --> Parameter containing:
tensor([0., 0.], requires_grad=True)
-----------------------打印字典-----------------------
attribute4cxq tensor(20200910.)
attribute4lzq tensor(20200.)
attribute4zh tensor([0., 0.])
attribute4cjhP tensor([6.5713e+05, 1.5286e-19])
登记注册attribute4cjhRP属性 tensor([0., 0.])
(ssd4pytorch1_2_0) PS C:\Users\chenxuqi\Desktop\News4cxq\test4cxq>