Trjtdtk

Can smartphones with the same camera sensor have different image quality?

Is above average number of years spent on PhD considered a red flag in future academia or industry positions?

Did the UK government pay "millions and millions of dollars" to try to snag Julian Assange?

How to pronounce 1ターン?

Take groceries in checked luggage

Difference between "generating set" and free product?

Does the AirPods case need to be around while listening via an iOS Device?

Why did all the guest students take carriages to the Yule Ball?

Why can't devices on different VLANs, but on the same subnet, communicate?

When did F become S in typeography, and why?

How did the audience guess the pentatonic scale in Bobby McFerrin's presentation?

Are spiders unable to hurt humans, especially very small spiders?

Windows 10: How to Lock (not sleep) laptop on lid close?

Does Parliament need to approve the new Brexit delay to 31 October 2019?

First use of “packing” as in carrying a gun

How is simplicity better than precision and clarity in prose?

Would it be possible to rearrange a dragon's flight muscle to somewhat circumvent the square-cube law?

Why does the Event Horizon Telescope (EHT) not include telescopes from Africa, Asia or Australia?

The variadic template constructor of my class cannot modify my class members, why is that so?

Would an alien lifeform be able to achieve space travel if lacking in vision?

Typeface like Times New Roman but with "tied" percent sign

What's the point in a preamp?

ELI5: Why do they say that Israel would have been the fourth country to land a spacecraft on the Moon and why do they call it low cost?

How does ice melt when immersed in water

Pytorch dynamic forward pass

1

$begingroup$

Does there exist a fast and convenient way for handling such a problem:

class MyModule(nn.Module):
def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

You see i want to access different layers in the network conditioned on an additional input, which is also a list. Further i want to process the whole batch at once and the output.shape != input.shape.

pytorch

share|improve this question

edited Mar 1 at 14:44

Andreas Look

asked Mar 1 at 14:37

Andreas LookAndreas Look

441111

$endgroup$

add a comment | 

1

$begingroup$

Does there exist a fast and convenient way for handling such a problem:

class MyModule(nn.Module):
def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

You see i want to access different layers in the network conditioned on an additional input, which is also a list. Further i want to process the whole batch at once and the output.shape != input.shape.

pytorch

share|improve this question

edited Mar 1 at 14:44

Andreas Look

asked Mar 1 at 14:37

Andreas LookAndreas Look

441111

$endgroup$

add a comment | 

$begingroup$

Does there exist a fast and convenient way for handling such a problem:

class MyModule(nn.Module):
def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

You see i want to access different layers in the network conditioned on an additional input, which is also a list. Further i want to process the whole batch at once and the output.shape != input.shape.

pytorch

share|improve this question

edited Mar 1 at 14:44

Andreas Look

asked Mar 1 at 14:37

Andreas LookAndreas Look

441111

$endgroup$

class MyModule(nn.Module):
def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

share|improve this question

edited Mar 1 at 14:44

Andreas Look

asked Mar 1 at 14:37

Andreas LookAndreas Look

441111

share|improve this question

edited Mar 1 at 14:44

Andreas Look

asked Mar 1 at 14:37

Andreas LookAndreas Look

441111

asked Mar 1 at 14:37

Andreas LookAndreas Look

441111

asked Mar 1 at 14:37

Andreas LookAndreas Look

441111

1 Answer
1

active

oldest

votes

0

$begingroup$

Here is my understanding of your problem:

# Import
from torch import nn

# Define custom class
class MyModule(nn.Module):
 def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

 def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

# Intialize custom class 
net = MyModule()

# Access networks layers based on additional input
additional_input = 1
if additional_input == 1:
 idx = 0
 print(net.linears[idx].in_features)

share|improve this answer

answered Mar 1 at 18:32

Brian SpieringBrian Spiering

4,2681129

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  thank you brian. i think i didnt specify my problem good enough. by accessing the layers i actually meant to make a forward pass for a whole batch at once, while i need different layers for each entry in the batch conditioned on the extra input
  $endgroup$
  – Andreas Look
  Mar 1 at 18:57
  

  
  
  
  

add a comment | 

Your Answer

StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "557"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);

else
createEditor();

);

function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);



);

draft saved

draft discarded

Sign up or log in

StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);

Sign up using Google

Sign up using Facebook

Sign up using Email and Password

Post as a guest

Name

Email

Required, but never shown

StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f46471%2fpytorch-dynamic-forward-pass%23new-answer', 'question_page');

);

Post as a guest

Name

Email

Required, but never shown

0

$begingroup$

Here is my understanding of your problem:

# Import
from torch import nn

# Define custom class
class MyModule(nn.Module):
 def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

 def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

# Intialize custom class 
net = MyModule()

# Access networks layers based on additional input
additional_input = 1
if additional_input == 1:
 idx = 0
 print(net.linears[idx].in_features)

share|improve this answer

answered Mar 1 at 18:32

Brian SpieringBrian Spiering

4,2681129

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  thank you brian. i think i didnt specify my problem good enough. by accessing the layers i actually meant to make a forward pass for a whole batch at once, while i need different layers for each entry in the batch conditioned on the extra input
  $endgroup$
  – Andreas Look
  Mar 1 at 18:57
  

  
  
  
  

add a comment | 

0

$begingroup$

Here is my understanding of your problem:

# Import
from torch import nn

# Define custom class
class MyModule(nn.Module):
 def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

 def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

# Intialize custom class 
net = MyModule()

# Access networks layers based on additional input
additional_input = 1
if additional_input == 1:
 idx = 0
 print(net.linears[idx].in_features)

share|improve this answer

answered Mar 1 at 18:32

Brian SpieringBrian Spiering

4,2681129

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  thank you brian. i think i didnt specify my problem good enough. by accessing the layers i actually meant to make a forward pass for a whole batch at once, while i need different layers for each entry in the batch conditioned on the extra input
  $endgroup$
  – Andreas Look
  Mar 1 at 18:57
  

  
  
  
  

add a comment | 

$begingroup$

Here is my understanding of your problem:

# Import
from torch import nn

# Define custom class
class MyModule(nn.Module):
 def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

 def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

# Intialize custom class 
net = MyModule()

# Access networks layers based on additional input
additional_input = 1
if additional_input == 1:
 idx = 0
 print(net.linears[idx].in_features)

share|improve this answer

answered Mar 1 at 18:32

Brian SpieringBrian Spiering

4,2681129

$endgroup$

# Import
from torch import nn

# Define custom class
class MyModule(nn.Module):
 def __init__(self):
 super(MyModule, self).__init__()
 self.linears = nn.ModuleList([nn.Linear(10, 20) for _ in range(10)])

 def forward(self, x, indices):
 x = self.linears[indices](x) 
 return x

# Intialize custom class 
net = MyModule()

# Access networks layers based on additional input
additional_input = 1
if additional_input == 1:
 idx = 0
 print(net.linears[idx].in_features)

share|improve this answer

answered Mar 1 at 18:32

Brian SpieringBrian Spiering

4,2681129

answered Mar 1 at 18:32

Brian SpieringBrian Spiering

4,2681129

answered Mar 1 at 18:32

Brian SpieringBrian Spiering

4,2681129