Overfit problem in siamese network with triplet loss The 2019 Stack Overflow Developer Survey Results Are InTrying to come up with a feature to improve emotion classifier based on facial movement using facial landmarksConvolutional Neural Network not learning EEG dataHow to know the model has started overfitting?Cannot underfit/overfit on the IRIS datasetHow to combine heterogeneous image features extracted with different algorithms for similar image retrieval?Convolution neural network with 11 million parameters unable to overfit on 100 image samplesOverfitting in Siamese NetworkOverfit from the loss curveArcFace loss in siamese architecture?Siamese network using VGG16 to verify the similarity images
Why didn't the Event Horizon Telescope team mention Sagittarius A*?
Are spiders unable to hurt humans, especially very small spiders?
Can you compress metal and what would be the consequences?
Command for nulifying spaces
How can I add encounters in the Lost Mine of Phandelver campaign without giving PCs too much XP?
Does a dangling wire really electrocute me if I'm standing in water?
Why doesn't mkfifo with a mode of 1755 grant read permissions and sticky bit to the user?
Is "plugging out" electronic devices an American expression?
Why is the maximum length of OpenWrt’s root password 8 characters?
Why not take a picture of a closer black hole?
Is there a better way to do an empty check in Java?
"consumers choosing to rely" vs. "consumers to choose to rely"
Is bread bad for ducks?
Why don't hard Brexiteers insist on a hard border to prevent illegal immigration after Brexit?
Did Scotland spend $250,000 for the slogan "Welcome to Scotland"?
How to display lines in a file like ls displays files in a directory?
Geography at the pixel level
Cooking pasta in a water boiler
How to type this arrow in math mode?
What to expect from an e-bike service?
Mathematics of imaging the black hole
What could be the right powersource for 15 seconds lifespan disposable giant chainsaw?
Old scifi movie from the 50s or 60s with men in solid red uniforms who interrogate a spy from the past
Landlord wants to switch my lease to a "Land contract" to "get back at the city"
Overfit problem in siamese network with triplet loss
The 2019 Stack Overflow Developer Survey Results Are InTrying to come up with a feature to improve emotion classifier based on facial movement using facial landmarksConvolutional Neural Network not learning EEG dataHow to know the model has started overfitting?Cannot underfit/overfit on the IRIS datasetHow to combine heterogeneous image features extracted with different algorithms for similar image retrieval?Convolution neural network with 11 million parameters unable to overfit on 100 image samplesOverfitting in Siamese NetworkOverfit from the loss curveArcFace loss in siamese architecture?Siamese network using VGG16 to verify the similarity images
$begingroup$
I have built a siamese network which uses triplet loss for heterogeneous face recognition system. I have created $40000$ triplets from $20$ subject each of them has $5$ anchors. Also note that I am using pretrained Resnet50 network as my base siamese networks. I am getting 99% training accuracy and $30%-50%$ test accuracy. How can i solve this overfitting problem for such a small dataset? Thank you in advance.
deep-learning convnet computer-vision overfitting
edited Mar 30 at 14:06
Siong Thye Goh
1,408620
asked Mar 30 at 13:26
Alperen KantarcıAlperen Kantarcı
1012
$endgroup$
add a comment |
$begingroup$
I have built a siamese network which uses triplet loss for heterogeneous face recognition system. I have created $40000$ triplets from $20$ subject each of them has $5$ anchors. Also note that I am using pretrained Resnet50 network as my base siamese networks. I am getting 99% training accuracy and $30%-50%$ test accuracy. How can i solve this overfitting problem for such a small dataset? Thank you in advance.
deep-learning convnet computer-vision overfitting
edited Mar 30 at 14:06
Siong Thye Goh
1,408620
asked Mar 30 at 13:26
Alperen KantarcıAlperen Kantarcı
1012
$endgroup$
add a comment |
$begingroup$
I have built a siamese network which uses triplet loss for heterogeneous face recognition system. I have created $40000$ triplets from $20$ subject each of them has $5$ anchors. Also note that I am using pretrained Resnet50 network as my base siamese networks. I am getting 99% training accuracy and $30%-50%$ test accuracy. How can i solve this overfitting problem for such a small dataset? Thank you in advance.
deep-learning convnet computer-vision overfitting
edited Mar 30 at 14:06
Siong Thye Goh
1,408620
asked Mar 30 at 13:26
Alperen KantarcıAlperen Kantarcı
1012
$endgroup$
I have built a siamese network which uses triplet loss for heterogeneous face recognition system. I have created $40000$ triplets from $20$ subject each of them has $5$ anchors. Also note that I am using pretrained Resnet50 network as my base siamese networks. I am getting 99% training accuracy and $30%-50%$ test accuracy. How can i solve this overfitting problem for such a small dataset? Thank you in advance.
deep-learning convnet computer-vision overfitting
deep-learning convnet computer-vision overfitting
edited Mar 30 at 14:06
Siong Thye Goh
1,408620
asked Mar 30 at 13:26
Alperen KantarcıAlperen Kantarcı
1012
edited Mar 30 at 14:06
Siong Thye Goh
1,408620
asked Mar 30 at 13:26
Alperen KantarcıAlperen Kantarcı
1012
edited Mar 30 at 14:06
Siong Thye Goh
1,408620
edited Mar 30 at 14:06
Siong Thye Goh
1,408620
edited Mar 30 at 14:06
Siong Thye Goh
1,408620
1,408620
asked Mar 30 at 13:26
Alperen KantarcıAlperen Kantarcı
1012
asked Mar 30 at 13:26
Alperen KantarcıAlperen Kantarcı
1012
asked Mar 30 at 13:26
Alperen KantarcıAlperen Kantarcı
1012
1012
add a comment |
add a comment |
0
active
oldest
votes
Your Answer
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%2f48260%2foverfit-problem-in-siamese-network-with-triplet-loss%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
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%2f48260%2foverfit-problem-in-siamese-network-with-triplet-loss%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
