In [1]: import torchvision
In [2]: import torch
In [3]: import torch.nn as nn
In [4]: alex = torchvision.models.alexnet(pretrained=False)
In [5]: alex?
Signature: alex(*input, **kwargs)
Type: AlexNet
String form:
AlexNet(
(features): Sequential(
(0): Conv2d(3, 64, kernel_size=(11, 11), stride=(4, 4), pa <...> rue)
(5): ReLU(inplace)
(6): Linear(in_features=4096, out_features=1000, bias=True)
)
)
File: ~/anaconda3/lib/python3.6/site-packages/torchvision/models/alexnet.py
Docstring: <no docstring>
Class docstring:
Base class for all neural network modules.
Your models should also subclass this class.
Modules can also contain other Modules, allowing to nest them in
a tree structure. You can assign the submodules as regular attributes::
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self):
super(Model, self).__init__()
self.conv1 = nn.Conv2d(1, 20, 5)
self.conv2 = nn.Conv2d(20, 20, 5)
def forward(self, x):
x = F.relu(self.conv1(x))
return F.relu(self.conv2(x))
Submodules assigned in this way will be registered, and will have their
parameters converted too when you call :meth:`to`, etc.
In [6]: torchvision.models.AlexNet?
Init signature: torchvision.models.AlexNet(num_classes=1000)
Docstring:
Base class for all neural network modules.
Your models should also subclass this class.
Modules can also contain other Modules, allowing to nest them in
a tree structure. You can assign the submodules as regular attributes::
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self):
super(Model, self).__init__()
self.conv1 = nn.Conv2d(1, 20, 5)
self.conv2 = nn.Conv2d(20, 20, 5)
def forward(self, x):
x = F.relu(self.conv1(x))
return F.relu(self.conv2(x))
Submodules assigned in this way will be registered, and will have their
parameters converted too when you call :meth:`to`, etc.
File: ~/anaconda3/lib/python3.6/site-packages/torchvision/models/alexnet.py
Type: type
In [7]: class realex(alex):
...: def __init__(self):
...: super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return F.relu(self.conv2(x))
...:
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-7-e07645723006> in <module>
----> 1 class realex(alex):
2 def __init__(self):
3 super().__init__()
4 self.conv1 = nn.Conv1d(1, 20, 5)
5 self.conv2 = nn.Conv2d(20, 20, 6)
TypeError: __init__() takes from 1 to 2 positional arguments but 4 were given
In [8]: class realex(nn.Module):
...: def __init__(self):
...: super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return F.relu(self.conv2(x))
...:
In [9]: class realex(torchvision.models.AlexNet):
...: def __init__(self):
...: super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return F.relu(self.conv2(x))
...:
In [10]: realex
Out[10]: __main__.realex
In [11]: realexx = realex()
In [12]: realexx
Out[12]:
realex(
(features): Sequential(
(0): Conv2d(3, 64, kernel_size=(11, 11), stride=(4, 4), padding=(2, 2))
(1): ReLU(inplace)
(2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
(3): Conv2d(64, 192, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(4): ReLU(inplace)
(5): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
(6): Conv2d(192, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(7): ReLU(inplace)
(8): Conv2d(384, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(9): ReLU(inplace)
(10): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(11): ReLU(inplace)
(12): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
)
(avgpool): AdaptiveAvgPool2d(output_size=(6, 6))
(classifier): Sequential(
(0): Dropout(p=0.5)
(1): Linear(in_features=9216, out_features=4096, bias=True)
(2): ReLU(inplace)
(3): Dropout(p=0.5)
(4): Linear(in_features=4096, out_features=4096, bias=True)
(5): ReLU(inplace)
(6): Linear(in_features=4096, out_features=1000, bias=True)
)
(conv1): Conv1d(1, 20, kernel_size=(5,), stride=(1,))
(conv2): Conv2d(20, 20, kernel_size=(6, 6), stride=(1, 1))
)
In [13]: class realex(torchvision.models.AlexNet):
...: def __init__(self):
...: #super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return F.relu(self.conv2(x))
...:
In [14]: realexx
Out[14]:
realex(
(features): Sequential(
(0): Conv2d(3, 64, kernel_size=(11, 11), stride=(4, 4), padding=(2, 2))
(1): ReLU(inplace)
(2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
(3): Conv2d(64, 192, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(4): ReLU(inplace)
(5): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
(6): Conv2d(192, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(7): ReLU(inplace)
(8): Conv2d(384, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(9): ReLU(inplace)
(10): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(11): ReLU(inplace)
(12): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
)
(avgpool): AdaptiveAvgPool2d(output_size=(6, 6))
(classifier): Sequential(
(0): Dropout(p=0.5)
(1): Linear(in_features=9216, out_features=4096, bias=True)
(2): ReLU(inplace)
(3): Dropout(p=0.5)
(4): Linear(in_features=4096, out_features=4096, bias=True)
(5): ReLU(inplace)
(6): Linear(in_features=4096, out_features=1000, bias=True)
)
(conv1): Conv1d(1, 20, kernel_size=(5,), stride=(1,))
(conv2): Conv2d(20, 20, kernel_size=(6, 6), stride=(1, 1))
)
In [15]: import os
In [16]: os.system('vim ~/anaconda3/lib/python3.6/site-packages/torchvision/models/alexnet.py')
Out[16]: 0
In [17]: class realex(torchvision.models.AlexNet):
...: def __init__(self):
...: super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return F.relu(self.conv2(x))
...:
In [18]: real = realex()
In [19]: real
Out[19]:
realex(
(features): Sequential(
(0): Conv2d(3, 64, kernel_size=(11, 11), stride=(4, 4), padding=(2, 2))
(1): ReLU(inplace)
(2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
(3): Conv2d(64, 192, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))
(4): ReLU(inplace)
(5): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
(6): Conv2d(192, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(7): ReLU(inplace)
(8): Conv2d(384, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(9): ReLU(inplace)
(10): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(11): ReLU(inplace)
(12): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=False)
)
(avgpool): AdaptiveAvgPool2d(output_size=(6, 6))
(classifier): Sequential(
(0): Dropout(p=0.5)
(1): Linear(in_features=9216, out_features=4096, bias=True)
(2): ReLU(inplace)
(3): Dropout(p=0.5)
(4): Linear(in_features=4096, out_features=4096, bias=True)
(5): ReLU(inplace)
(6): Linear(in_features=4096, out_features=1000, bias=True)
)
(conv1): Conv1d(1, 20, kernel_size=(5,), stride=(1,))
(conv2): Conv2d(20, 20, kernel_size=(6, 6), stride=(1, 1))
)
In [20]: class realex(nn.Module):
...: def __init__(self):
...: super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return F.relu(self.conv2(x))
...:
In [21]: real = realex()
In [22]: real
Out[22]:
realex(
(conv1): Conv1d(1, 20, kernel_size=(5,), stride=(1,))
(conv2): Conv2d(20, 20, kernel_size=(6, 6), stride=(1, 1))
)
In [23]: nn.Module?
Init signature: nn.Module()
Docstring:
Base class for all neural network modules.
Your models should also subclass this class.
Modules can also contain other Modules, allowing to nest them in
a tree structure. You can assign the submodules as regular attributes::
import torch.nn as nn
import torch.nn.functional as F
class Model(nn.Module):
def __init__(self):
super(Model, self).__init__()
self.conv1 = nn.Conv2d(1, 20, 5)
self.conv2 = nn.Conv2d(20, 20, 5)
def forward(self, x):
x = F.relu(self.conv1(x))
return F.relu(self.conv2(x))
Submodules assigned in this way will be registered, and will have their
parameters converted too when you call :meth:`to`, etc.
File: ~/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py
Type: type
In [24]: os.system('vim ~/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py')
Out[24]: 0
In [25]: class realex(nn.Module):
...: def __init__(self):
...: super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return x
...:
In [26]: real = realex()
In [27]: real
Out[27]:
realex(
(conv1): Conv1d(1, 20, kernel_size=(5,), stride=(1,))
(conv2): Conv2d(20, 20, kernel_size=(6, 6), stride=(1, 1))
)
In [28]: class realex(nn.Module):
...: def __init__(self):
...: super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: nn.ReLU(inplace=True)
...: nn.Dropout(inplace=True)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return x
...:
In [29]: rea = realex()
In [30]: rea
Out[30]:
realex(
(conv1): Conv1d(1, 20, kernel_size=(5,), stride=(1,))
(conv2): Conv2d(20, 20, kernel_size=(6, 6), stride=(1, 1))
)
In [31]: class realex(nn.Module):
...: def __init__(self):
...: super().__init__()
...: self.conv1 = nn.Conv1d(1, 20, 5)
...: self.relu = nn.ReLU(inplace=True)
...: nn.Dropout(inplace=True)
...: self.conv2 = nn.Conv2d(20, 20, 6)
...: def forward(self, x):
...: x = F.relu(self.conv1(x))
...: return x
...:
In [32]: rea = realex()
In [33]: rea
Out[33]:
realex(
(conv1): Conv1d(1, 20, kernel_size=(5,), stride=(1,))
(relu): ReLU(inplace)
(conv2): Conv2d(20, 20, kernel_size=(6, 6), stride=(1, 1))
)
pytorch笔记
最新推荐文章于 2024-09-04 08:22:11 发布