import torch
import torchkeras
print("torch.__version__="+torch.__version__)
print("torchkeras.__version__="+torchkeras.__version__)
"""
torch.__version__=2.1.1+cu118
torchkeras.__version__=3.9.4
"""
1.nn.functional和nn.Module
前面我们介绍了Pytorch的张量的结构操作和数学运算中的一些常用API。
利用这些张量的API我们可以构建出神经网络相关的组件(如激活函数,模型层,损失函数)
Pytorch和神经网络相关的功能组件大多都封装在torch.nn模块下。
这些功能组件的绝大部分既有函数形式实现,也有类形式实现。
其中nn.functional(一般引入后改名为F)有各种功能组件的函数实现,例如:
激活函数:
- F.relu
- F.sigmoid
- F.tanh
- F.softmax
模型层:
- F.linear
- F.conv2d
- F.max_pool2d
- F.droupout2d
- F.embedding
损失函数
- F.binary_cross_entropy
- F.mse_loss
- F.cross_entropy
为了便于对参数进行管理,一般通过继承nn.Module转换成类的实现形式,并直接封装在nn模块下。例如:
激活函数
- nn.ReLU
- nn.Sigmoid
- nn.Tanh
- nn.Softmax
模型层
- nn.Linear
- nn.Conv2d
- nn.MaxPool2d
- nn.Dropout2d
- nn.Embedding
损失函数
- nn.BECLoss
- nn.MSELoss
- nn.CrossEntropyLoss
实际上nn.Module除了可以管理其引用的各种参数,还可以管理其引用的子模块,功能十分强大。
import torch
torch.relu(torch.tensor(-1.0))
"""
tensor(0.)
"""
import torch.nn.functional as F
F.relu(torch.tensor(-1.0))
"""
tensor(0.)
"""
2.使用nn.Module来管理参数(配合nn.Parameter使用)
在Pytorch中,模型的参数是需要被优化器训练的,因此,通常要设置参数为requires_grad=True的张量。
同时,在一个模型中,往往有许多的参数,要手动管理这些参数并不是一件容易的事情。
Pytorch一般将参数用nn.Parameter来表示,并且用nn.Module来管理其结构下的所有参数。
import torch
from torch import nn
import torch.nn.functional as F
torch.randn(2, 2, requires_grad=True)
"""
tensor([[-0.1423, -0.3451],
[ 0.1613, 1.0027]], requires_grad=True)
"""
# nn.Parameter具有requires_grad属性
w = nn.Parameter(torch.randn(2, 2))
print(w)
print(w.requires_grad)
"""
Parameter containing:
tensor([[ 0.1437, -1.5047],
[-0.0155, -0.3239]], requires_grad=True)
True
"""
# nn、ParameterList可以将多个nn.Parameter组成一个列表
params_list = nn.ParameterList([nn.Parameter(torch.rand(8, i)) for i in range(1, 3)])
print(params_list)
print(params_list[0].requires_grad)
"""
ParameterList(
(0): Parameter containing: [torch.float32 of size 8x1]
(1): Parameter containing: [torch.float32 of size 8x2]
)
True
"""
# nn.ParameterDict可以将多个nn.Parameter组成一个字典
param_dict = nn.ParameterDict({
'a': nn.Parameter(torch.rand(2, 2)),
'b': nn.Parameter(torch.zeros(2))
})
print(param_dict)
print(param_dict['a'].requires_grad)
"""
ParameterDict(
(a): Parameter containing: [torch.FloatTensor of size 2x2]
(b): Parameter containing: [torch.FloatTensor of size 2]
)
True
"""
# 可以用Module将它们管理起来
# module.parameters()返回一个生成器,包括其结构下的所有parameters
module = nn.Module()
module.w = nn.Parameter(torch.randn(2, 2))
module.params_list = nn.ParameterList([nn.Parameter(torch.rand(8,i)) for i in range(1,3)])
module.params_dict = nn.ParameterDict({
'a': nn.Parameter(torch.rand(2, 2)),
'b': nn.Parameter(torch.zeros(2))
})
num_param = 0
for param in module.named_parameters():
print(param)
num_param = num_param + 1
print('number of Parameters=', num_param)
"""
('w', Parameter containing:
tensor([[-0.9218, -0.0231],
[-0.6829, 0.9799]], requires_grad=True))
('params_list.0', Parameter containing:
tensor([[0.6197],
[0.4622],
[0.8464],
[0.9107],
[0.3164],
[0.0340],
[0.0739],
[0.4249]], requires_grad=True))
('params_list.1', Parameter containing:
tensor([[0.9176, 0.5050],
[0.2121, 0.7156],
[0.2281, 0.5439],
[0.5997, 0.7202],
[0.9591, 0.6104],
[0.5161, 0.1679],
[0.8338, 0.5305],
[0.2937, 0.0653]], requires_grad=True))
('params_dict.a', Parameter containing:
tensor([[0.7809, 0.4616],
[0.2082, 0.4271]], requires_grad=True))
('params_dict.b', Parameter containing:
tensor([0., 0.], requires_grad=True))
number of Parameters= 5
"""
# 实践当中,一般通过继承nn.Module来构建模块类,并将所有含有需要学习的参数的部分放在构造函数中
# 以下范例为Pytorch中nn.Linear的源码的简化版本
# 可以看到它将需要学习的参数放在了__init__构造函数中,并在forward中调用F.linear函数来实现计算逻辑
class Linear(nn.Module):
__constants__ = ['in_features', 'out_features']
def __init__(self, infeatures, out_features, bias=True):
super().__init__()
self.in_features = in_features
self.out_features = out_features
self.weight = nn.Parameter(torch.Tensor(out_features, in_features))
if bias:
self.bias = nn.Parameter(torch.Tensor(out_features))
else:
self.register_parameter('bias', None)
def forward(self, input):
return F.linear(input, self.weight, self.bias)
3.使用nn.Module来管理子模块
一般情况下,我们都很少直接使用nn.Parameter来定义参数构建模型,而是通过拼装一些常用的模型层来构造模型。
这些模型层也是继承自nn.Module的对象,本身也有包括参数,属于我们要定义的模块的子模块。
nn.Module提供了一些方法可以管理这些子模块。
- children()方法:返回生成器,包括模块下的所有子模块
- named_children()方法:返回一个生成器,包括模块下的所有子模块,以及它们的名字
- modules()方法:返回一个生成器,包括模块下的所有各个层级的模块,包括模块本身
- named_modules()方法:返回一个生成器,包括模块下的所有各个层级的模块以及它们的名字,包括模块本身
其中children()方法和named_children()方法较多使用。
modules()方法和named_modules()方法较少使用,其功能可以通过多个named_children()的嵌套使用实现。
class Net(nn.Module):
def __init__(self):
super().__init__()
self.embedding = nn.Embedding(num_embeddings=10000, embedding_dim=3, padding_idx=1)
self.conv = nn.Sequential()
self.conv.add_module('conv_1', nn.Conv1d(in_channels=3, out_channels=16, kernel_size=5))
self.conv.add_module('pool_1', nn.MaxPool1d(kernel_size=2))
self.conv.add_module('relu_1', nn.ReLU())
self.conv.add_module('conv_2', nn.Conv1d(in_channels=16, out_channels=128, kernel_size=2))
self.conv.add_module('pool_2', nn.MaxPool1d(kernel_size=2))
self.conv.add_module('relu_2', nn.ReLU())
self.dense = nn.Sequential()
self.dense.add_module('flatten', nn.Flatten())
self.dense.add_module('linear', nn.Linear(6144, 1))
def forward(self, x):
x = self.embedding(x).transpose(1, 2)
x = self.conv(x)
y = self.dense(x)
return y
net = Net()
i = 0
for child in net.children():
i += 1
print(child, '\n')
print('child number', i)
"""
Embedding(10000, 3, padding_idx=1)
Sequential(
(conv_1): Conv1d(3, 16, kernel_size=(5,), stride=(1,))
(pool_1): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_1): ReLU()
(conv_2): Conv1d(16, 128, kernel_size=(2,), stride=(1,))
(pool_2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_2): ReLU()
)
Sequential(
(flatten): Flatten(start_dim=1, end_dim=-1)
(linear): Linear(in_features=6144, out_features=1, bias=True)
)
child number 3
Selection deleted
"""
i = 0
for name, child in net.named_children():
i += 1
print(name, ':', child, '\n')
print('child number', i)
"""
embedding : Embedding(10000, 3, padding_idx=1)
conv : Sequential(
(conv_1): Conv1d(3, 16, kernel_size=(5,), stride=(1,))
(pool_1): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_1): ReLU()
(conv_2): Conv1d(16, 128, kernel_size=(2,), stride=(1,))
(pool_2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_2): ReLU()
)
dense : Sequential(
(flatten): Flatten(start_dim=1, end_dim=-1)
(linear): Linear(in_features=6144, out_features=1, bias=True)
)
child number 3
"""
i = 0
for module in net.modules():
i += 1
print(module)
print('module number:', i)
"""
Net(
(embedding): Embedding(10000, 3, padding_idx=1)
(conv): Sequential(
(conv_1): Conv1d(3, 16, kernel_size=(5,), stride=(1,))
(pool_1): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_1): ReLU()
(conv_2): Conv1d(16, 128, kernel_size=(2,), stride=(1,))
(pool_2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_2): ReLU()
)
(dense): Sequential(
(flatten): Flatten(start_dim=1, end_dim=-1)
(linear): Linear(in_features=6144, out_features=1, bias=True)
)
)
Embedding(10000, 3, padding_idx=1)
Sequential(
(conv_1): Conv1d(3, 16, kernel_size=(5,), stride=(1,))
(pool_1): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_1): ReLU()
(conv_2): Conv1d(16, 128, kernel_size=(2,), stride=(1,))
(pool_2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_2): ReLU()
)
Conv1d(3, 16, kernel_size=(5,), stride=(1,))
MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
ReLU()
Conv1d(16, 128, kernel_size=(2,), stride=(1,))
MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
ReLU()
Sequential(
(flatten): Flatten(start_dim=1, end_dim=-1)
(linear): Linear(in_features=6144, out_features=1, bias=True)
)
Flatten(start_dim=1, end_dim=-1)
Linear(in_features=6144, out_features=1, bias=True)
module number: 12
"""
下面我们通过named_children方法找到embedding层,并将其参数设置为不可训练(相当于冻结embedding层)
children_dict = {
name: module for name, module in net.named_children()
}
print(children_dict)
embedding = children_dict['embedding']
embedding.requires_grad_(False) # 冻结其参数
"""
{'embedding': Embedding(10000, 3, padding_idx=1), 'conv': Sequential(
(conv_1): Conv1d(3, 16, kernel_size=(5,), stride=(1,))
(pool_1): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_1): ReLU()
(conv_2): Conv1d(16, 128, kernel_size=(2,), stride=(1,))
(pool_2): MaxPool1d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(relu_2): ReLU()
), 'dense': Sequential(
(flatten): Flatten(start_dim=1, end_dim=-1)
(linear): Linear(in_features=6144, out_features=1, bias=True)
)}
Embedding(10000, 3, padding_idx=1)
"""
# 可以看到其第一层的参数已经不可以被训练了
for param in embedding.parameters():
print(param.requires_grad)
print(param.numel())
"""
False
30000
"""
from torchkeras import summary
summary(net, input_shape=(200, ), input_dtype=torch.LongTensor);
"""
--------------------------------------------------------------------------
Layer (type) Output Shape Param #
==========================================================================
Embedding-1 [-1, 200, 3] 30,000
Conv1d-2 [-1, 16, 196] 256
MaxPool1d-3 [-1, 16, 98] 0
ReLU-4 [-1, 16, 98] 0
Conv1d-5 [-1, 128, 97] 4,224
MaxPool1d-6 [-1, 128, 48] 0
ReLU-7 [-1, 128, 48] 0
Flatten-8 [-1, 6144] 0
Linear-9 [-1, 1] 6,145
==========================================================================
Total params: 40,625
Trainable params: 10,625
Non-trainable params: 30,000
--------------------------------------------------------------------------
Input size (MB): 0.000763
Forward/backward pass size (MB): 0.287788
Params size (MB): 0.154972
Estimated Total Size (MB): 0.443523
--------------------------------------------------------------------------
"""
标签:kernel,nn,3nn,self,torch,functional,Module,Parameter,size
From: https://www.cnblogs.com/lotuslaw/p/18064490