Why do I need pre-trained weights in transfer learning? The Next CEO of Stack Overflow2019 Community Moderator ElectionRight way to Fine Tune - Train a fully connected layer as a separate stepWhy is vanishing gradient a problem?What is the different between Fine-tuning and Transfer-learning?Fine tuning accuracy lower than Raw Transfer Learning AccuracyType of images used to train a neural networkWhy is input preprocessing in VGG16 in Keras not 1/255.0How to properly mean-center my dataWhy is the GoogLeNet retrained model size less compared to others?How to input different sized images into transfer learning networkOver fitting in Transfer Learning with small dataset
Aggressive Under-Indexing and no data for missing index
What flight has the highest ratio of timezone difference to flight time?
Which one is the true statement?
Is there an equivalent of cd - for cp or mv
Can I calculate next year's exemptions based on this year's refund/amount owed?
Is there a reasonable and studied concept of reduction between regular languages?
Purpose of level-shifter with same in and out voltages
Point distance program written without a framework
Film where the government was corrupt with aliens, people sent to kill aliens are given rigged visors not showing the right aliens
Calculate the Mean mean of two numbers
What CSS properties can the br tag have?
free fall ellipse or parabola?
Getting Stale Gas Out of a Gas Tank w/out Dropping the Tank
Do scriptures give a method to recognize a truly self-realized person/jivanmukta?
Traveling with my 5 year old daughter (as the father) without the mother from Germany to Mexico
Is it convenient to ask the journal's editor for two additional days to complete a review?
TikZ: How to fill area with a special pattern?
IC has pull-down resistors on SMBus lines?
What is the difference between "hamstring tendon" and "common hamstring tendon"?
What would be the main consequences for a country leaving the WTO?
Players Circumventing the limitations of Wish
Can Sneak Attack be used when hitting with an improvised weapon?
Expressing the idea of having a very busy time
Help understanding this unsettling image of Titan, Epimetheus, and Saturn's rings?
Why do I need pre-trained weights in transfer learning?
The Next CEO of Stack Overflow2019 Community Moderator ElectionRight way to Fine Tune - Train a fully connected layer as a separate stepWhy is vanishing gradient a problem?What is the different between Fine-tuning and Transfer-learning?Fine tuning accuracy lower than Raw Transfer Learning AccuracyType of images used to train a neural networkWhy is input preprocessing in VGG16 in Keras not 1/255.0How to properly mean-center my dataWhy is the GoogLeNet retrained model size less compared to others?How to input different sized images into transfer learning networkOver fitting in Transfer Learning with small dataset
$begingroup$
I am using a Mask-RCNN. I first chose the resnet50 backbone then downloaded COCO pre-trained weights.
Why do I need pre-trained weights for transfer learning?
The transfer learning approach is to train a base network and then copy its first layers of the target. My base network is resnet50 and I copied the first layers to Mask RCNN. So, why do I need pre-trained weights, for example, coco pre-trained weights?
I understand that I copy layers and pre-trained weights from resnet.
Thanks in advance.
deep-learning cnn transfer-learning
edited 3 mins ago
Ethan
612324
asked 5 hours ago
BadumBadum
114
$endgroup$
add a comment |
$begingroup$
I am using a Mask-RCNN. I first chose the resnet50 backbone then downloaded COCO pre-trained weights.
Why do I need pre-trained weights for transfer learning?
The transfer learning approach is to train a base network and then copy its first layers of the target. My base network is resnet50 and I copied the first layers to Mask RCNN. So, why do I need pre-trained weights, for example, coco pre-trained weights?
I understand that I copy layers and pre-trained weights from resnet.
Thanks in advance.
deep-learning cnn transfer-learning
edited 3 mins ago
Ethan
612324
asked 5 hours ago
BadumBadum
114
$endgroup$
add a comment |
$begingroup$
I am using a Mask-RCNN. I first chose the resnet50 backbone then downloaded COCO pre-trained weights.
Why do I need pre-trained weights for transfer learning?
The transfer learning approach is to train a base network and then copy its first layers of the target. My base network is resnet50 and I copied the first layers to Mask RCNN. So, why do I need pre-trained weights, for example, coco pre-trained weights?
I understand that I copy layers and pre-trained weights from resnet.
Thanks in advance.
deep-learning cnn transfer-learning
edited 3 mins ago
Ethan
612324
asked 5 hours ago
BadumBadum
114
$endgroup$
I am using a Mask-RCNN. I first chose the resnet50 backbone then downloaded COCO pre-trained weights.
Why do I need pre-trained weights for transfer learning?
The transfer learning approach is to train a base network and then copy its first layers of the target. My base network is resnet50 and I copied the first layers to Mask RCNN. So, why do I need pre-trained weights, for example, coco pre-trained weights?
I understand that I copy layers and pre-trained weights from resnet.
Thanks in advance.
deep-learning cnn transfer-learning
deep-learning cnn transfer-learning
edited 3 mins ago
Ethan
612324
asked 5 hours ago
BadumBadum
114
edited 3 mins ago
Ethan
612324
asked 5 hours ago
BadumBadum
114
edited 3 mins ago
Ethan
612324
edited 3 mins ago
Ethan
612324
edited 3 mins ago
Ethan
612324
612324
asked 5 hours ago
BadumBadum
114
asked 5 hours ago
BadumBadum
114
asked 5 hours ago
BadumBadum
114
114
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
$begingroup$
You need pre-trainned weights for it to be Transfer Learning.
Copying layer structures is not transfer learning, it is just structuring a network inspired/copied from others.
The transferm learning lies in using pre-trained layers to construct a different network that migth have similarities in the first layers. That is usually useful for Deep Learning.
So the why's to use Transfer Learning are:
You want to analyse something different in a dataset that was used to train another network
You want to perform classification in a class that was used to train a certain network but was not annotated before
You want to train a network for a problem similar to the other one and don't have time or computational power to retrain all layers
answered 1 hour ago
Pedro Henrique MonfortePedro Henrique Monforte
905
$endgroup$
add a comment |
StackExchange.ifUsing("editor", function ()
return StackExchange.using("mathjaxEditing", function ()
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
);
);
, "mathjax-editing");
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
);
);
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%2fdatascience.stackexchange.com%2fquestions%2f48317%2fwhy-do-i-need-pre-trained-weights-in-transfer-learning%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
$begingroup$
You need pre-trainned weights for it to be Transfer Learning.
Copying layer structures is not transfer learning, it is just structuring a network inspired/copied from others.
The transferm learning lies in using pre-trained layers to construct a different network that migth have similarities in the first layers. That is usually useful for Deep Learning.
So the why's to use Transfer Learning are:
You want to analyse something different in a dataset that was used to train another network
You want to perform classification in a class that was used to train a certain network but was not annotated before
You want to train a network for a problem similar to the other one and don't have time or computational power to retrain all layers
answered 1 hour ago
Pedro Henrique MonfortePedro Henrique Monforte
905
$endgroup$
add a comment |
$begingroup$
You need pre-trainned weights for it to be Transfer Learning.
Copying layer structures is not transfer learning, it is just structuring a network inspired/copied from others.
The transferm learning lies in using pre-trained layers to construct a different network that migth have similarities in the first layers. That is usually useful for Deep Learning.
So the why's to use Transfer Learning are:
You want to analyse something different in a dataset that was used to train another network
You want to perform classification in a class that was used to train a certain network but was not annotated before
You want to train a network for a problem similar to the other one and don't have time or computational power to retrain all layers
answered 1 hour ago
Pedro Henrique MonfortePedro Henrique Monforte
905
$endgroup$
add a comment |
$begingroup$
You need pre-trainned weights for it to be Transfer Learning.
Copying layer structures is not transfer learning, it is just structuring a network inspired/copied from others.
The transferm learning lies in using pre-trained layers to construct a different network that migth have similarities in the first layers. That is usually useful for Deep Learning.
So the why's to use Transfer Learning are:
You want to analyse something different in a dataset that was used to train another network
You want to perform classification in a class that was used to train a certain network but was not annotated before
You want to train a network for a problem similar to the other one and don't have time or computational power to retrain all layers
answered 1 hour ago
Pedro Henrique MonfortePedro Henrique Monforte
905
$endgroup$
You need pre-trainned weights for it to be Transfer Learning.
Copying layer structures is not transfer learning, it is just structuring a network inspired/copied from others.
The transferm learning lies in using pre-trained layers to construct a different network that migth have similarities in the first layers. That is usually useful for Deep Learning.
So the why's to use Transfer Learning are:
You want to analyse something different in a dataset that was used to train another network
You want to perform classification in a class that was used to train a certain network but was not annotated before
You want to train a network for a problem similar to the other one and don't have time or computational power to retrain all layers
answered 1 hour ago
Pedro Henrique MonfortePedro Henrique Monforte
905
answered 1 hour ago
Pedro Henrique MonfortePedro Henrique Monforte
905
answered 1 hour ago
Pedro Henrique MonfortePedro Henrique Monforte
905
answered 1 hour ago
Pedro Henrique MonfortePedro Henrique Monforte
905
905
add a comment |
add a comment |
Thanks for contributing an answer to Data Science Stack Exchange!
- 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.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
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%2fdatascience.stackexchange.com%2fquestions%2f48317%2fwhy-do-i-need-pre-trained-weights-in-transfer-learning%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
