Does a maxpooling layer reduce the number of parameters in a network?
0
I have a simple network defined:
model = Sequential()
model.add(Conv1D(5, 3, activation='relu', input_shape=(10, 1),name="conv1",padding="same"))
model.add(MaxPooling1D())
model.add(Conv1D(5, 3, activation='relu',name="conv2",padding="same"))
model.add(MaxPooling1D())
model.add(Dense(1, activation='relu',name="dense1"))
model.compile(loss='mse', optimizer='rmsprop')
The shape of the layers is as follows:
conv1-(None, 10, 5)
max1-(None, 5, 5)
conv2-(None,5,5)
max2-(None,2,5)
dense1-(None,2,1)
The model has a total of 106 parameters, however if I remove max pooling layer then the model summary looks as follows:
conv1-(None, 10, 5)
conv2-(None,10,5)
dense1-(None,10,1)
In both the cases total parameters remain 106, but why is it commonly written that the max-pooling layer reduces the number of parameters?
keras conv-neural-network max-pooling
edited Nov 26 '18 at 17:21
Daniel Möller
35.7k667104
asked Nov 26 '18 at 17:14
vampiretapvampiretap
565
add a comment |
0
I have a simple network defined:
model = Sequential()
model.add(Conv1D(5, 3, activation='relu', input_shape=(10, 1),name="conv1",padding="same"))
model.add(MaxPooling1D())
model.add(Conv1D(5, 3, activation='relu',name="conv2",padding="same"))
model.add(MaxPooling1D())
model.add(Dense(1, activation='relu',name="dense1"))
model.compile(loss='mse', optimizer='rmsprop')
The shape of the layers is as follows:
conv1-(None, 10, 5)
max1-(None, 5, 5)
conv2-(None,5,5)
max2-(None,2,5)
dense1-(None,2,1)
The model has a total of 106 parameters, however if I remove max pooling layer then the model summary looks as follows:
conv1-(None, 10, 5)
conv2-(None,10,5)
dense1-(None,10,1)
In both the cases total parameters remain 106, but why is it commonly written that the max-pooling layer reduces the number of parameters?
keras conv-neural-network max-pooling
edited Nov 26 '18 at 17:21
Daniel Möller
35.7k667104
asked Nov 26 '18 at 17:14
vampiretapvampiretap
565
add a comment |
0
0
0
I have a simple network defined:
model = Sequential()
model.add(Conv1D(5, 3, activation='relu', input_shape=(10, 1),name="conv1",padding="same"))
model.add(MaxPooling1D())
model.add(Conv1D(5, 3, activation='relu',name="conv2",padding="same"))
model.add(MaxPooling1D())
model.add(Dense(1, activation='relu',name="dense1"))
model.compile(loss='mse', optimizer='rmsprop')
The shape of the layers is as follows:
conv1-(None, 10, 5)
max1-(None, 5, 5)
conv2-(None,5,5)
max2-(None,2,5)
dense1-(None,2,1)
The model has a total of 106 parameters, however if I remove max pooling layer then the model summary looks as follows:
conv1-(None, 10, 5)
conv2-(None,10,5)
dense1-(None,10,1)
In both the cases total parameters remain 106, but why is it commonly written that the max-pooling layer reduces the number of parameters?
keras conv-neural-network max-pooling
edited Nov 26 '18 at 17:21
Daniel Möller
35.7k667104
asked Nov 26 '18 at 17:14
vampiretapvampiretap
565
I have a simple network defined:
model = Sequential()
model.add(Conv1D(5, 3, activation='relu', input_shape=(10, 1),name="conv1",padding="same"))
model.add(MaxPooling1D())
model.add(Conv1D(5, 3, activation='relu',name="conv2",padding="same"))
model.add(MaxPooling1D())
model.add(Dense(1, activation='relu',name="dense1"))
model.compile(loss='mse', optimizer='rmsprop')
The shape of the layers is as follows:
conv1-(None, 10, 5)
max1-(None, 5, 5)
conv2-(None,5,5)
max2-(None,2,5)
dense1-(None,2,1)
The model has a total of 106 parameters, however if I remove max pooling layer then the model summary looks as follows:
conv1-(None, 10, 5)
conv2-(None,10,5)
dense1-(None,10,1)
In both the cases total parameters remain 106, but why is it commonly written that the max-pooling layer reduces the number of parameters?
keras conv-neural-network max-pooling
keras conv-neural-network max-pooling
edited Nov 26 '18 at 17:21
Daniel Möller
35.7k667104
asked Nov 26 '18 at 17:14
vampiretapvampiretap
565
edited Nov 26 '18 at 17:21
Daniel Möller
35.7k667104
asked Nov 26 '18 at 17:14
vampiretapvampiretap
565
edited Nov 26 '18 at 17:21
Daniel Möller
35.7k667104
edited Nov 26 '18 at 17:21
Daniel Möller
35.7k667104
edited Nov 26 '18 at 17:21
Daniel Möller
35.7k667104
35.7k667104
asked Nov 26 '18 at 17:14
vampiretapvampiretap
565
asked Nov 26 '18 at 17:14
vampiretapvampiretap
565
asked Nov 26 '18 at 17:14
vampiretapvampiretap
565
565
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
0
Which kind of network? It's all up to you.
- Conv layers: no
- Dense layers:
- Directly after Conv or Pooling:
- With "channels_last": no
- With "channels_first": yes
- After Flatten layers: yes
- After GlobalPooling layers: no
- Directly after Conv or Pooling:
Your network: no.
Explanations
- Poolings and GlobalPoolings change the image sizes, but don't change the number of channels
- Conv layers are fixed size filters that stride along the images.The filter size is independent of the image size, thus there is no change. Filters depend on kernel size and channels
- Dense layers work on the last dimension only.
- If the last dimension is channels, the pooling layers don't affect it
- If the last dimension is an image side, it's affected
- Flatten layers transform the image sizes and channels into a single dimension.
answered Nov 26 '18 at 17:20
Daniel MöllerDaniel Möller
35.7k667104
add a comment |
Your Answer
StackExchange.ifUsing("editor", function () {
StackExchange.using("externalEditor", function () {
StackExchange.using("snippets", function () {
StackExchange.snippets.init();
});
});
}, "code-snippets");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "1"
};
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: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
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
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53486016%2fdoes-a-maxpooling-layer-reduce-the-number-of-parameters-in-a-network%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
0
Which kind of network? It's all up to you.
- Conv layers: no
- Dense layers:
- Directly after Conv or Pooling:
- With "channels_last": no
- With "channels_first": yes
- After Flatten layers: yes
- After GlobalPooling layers: no
- Directly after Conv or Pooling:
Your network: no.
Explanations
- Poolings and GlobalPoolings change the image sizes, but don't change the number of channels
- Conv layers are fixed size filters that stride along the images.The filter size is independent of the image size, thus there is no change. Filters depend on kernel size and channels
- Dense layers work on the last dimension only.
- If the last dimension is channels, the pooling layers don't affect it
- If the last dimension is an image side, it's affected
- Flatten layers transform the image sizes and channels into a single dimension.
answered Nov 26 '18 at 17:20
Daniel MöllerDaniel Möller
35.7k667104
add a comment |
0
Which kind of network? It's all up to you.
- Conv layers: no
- Dense layers:
- Directly after Conv or Pooling:
- With "channels_last": no
- With "channels_first": yes
- After Flatten layers: yes
- After GlobalPooling layers: no
- Directly after Conv or Pooling:
Your network: no.
Explanations
- Poolings and GlobalPoolings change the image sizes, but don't change the number of channels
- Conv layers are fixed size filters that stride along the images.The filter size is independent of the image size, thus there is no change. Filters depend on kernel size and channels
- Dense layers work on the last dimension only.
- If the last dimension is channels, the pooling layers don't affect it
- If the last dimension is an image side, it's affected
- Flatten layers transform the image sizes and channels into a single dimension.
answered Nov 26 '18 at 17:20
Daniel MöllerDaniel Möller
35.7k667104
add a comment |
0
0
0
Which kind of network? It's all up to you.
- Conv layers: no
- Dense layers:
- Directly after Conv or Pooling:
- With "channels_last": no
- With "channels_first": yes
- After Flatten layers: yes
- After GlobalPooling layers: no
- Directly after Conv or Pooling:
Your network: no.
Explanations
- Poolings and GlobalPoolings change the image sizes, but don't change the number of channels
- Conv layers are fixed size filters that stride along the images.The filter size is independent of the image size, thus there is no change. Filters depend on kernel size and channels
- Dense layers work on the last dimension only.
- If the last dimension is channels, the pooling layers don't affect it
- If the last dimension is an image side, it's affected
- Flatten layers transform the image sizes and channels into a single dimension.
answered Nov 26 '18 at 17:20
Daniel MöllerDaniel Möller
35.7k667104
Which kind of network? It's all up to you.
- Conv layers: no
- Dense layers:
- Directly after Conv or Pooling:
- With "channels_last": no
- With "channels_first": yes
- After Flatten layers: yes
- After GlobalPooling layers: no
- Directly after Conv or Pooling:
Your network: no.
Explanations
- Poolings and GlobalPoolings change the image sizes, but don't change the number of channels
- Conv layers are fixed size filters that stride along the images.The filter size is independent of the image size, thus there is no change. Filters depend on kernel size and channels
- Dense layers work on the last dimension only.
- If the last dimension is channels, the pooling layers don't affect it
- If the last dimension is an image side, it's affected
- Flatten layers transform the image sizes and channels into a single dimension.
answered Nov 26 '18 at 17:20
Daniel MöllerDaniel Möller
35.7k667104
answered Nov 26 '18 at 17:20
Daniel MöllerDaniel Möller
35.7k667104
answered Nov 26 '18 at 17:20
Daniel MöllerDaniel Möller
35.7k667104
answered Nov 26 '18 at 17:20
Daniel MöllerDaniel Möller
35.7k667104
35.7k667104
add a comment |
add a comment |
draft saved
draft discarded
Thanks for contributing an answer to Stack Overflow!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
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
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fstackoverflow.com%2fquestions%2f53486016%2fdoes-a-maxpooling-layer-reduce-the-number-of-parameters-in-a-network%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
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
Required, but never shown
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
Required, but never shown
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
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
