keras: how to add weights to loss evaluation
0
Todo :
- I would like to add a weight for each pattern loss in a given Keras loss function.
For example:
if the error on pattern i is l_i,
I would like to consider, instead, an error l_i * c_i, where c_i is an input scalar.
keras
edited Nov 13 '18 at 10:36
Abhishek Kumar
1,020216
asked Nov 13 '18 at 10:27
Filippo PorteraFilippo Portera
206
add a comment |
0
Todo :
- I would like to add a weight for each pattern loss in a given Keras loss function.
For example:
if the error on pattern i is l_i,
I would like to consider, instead, an error l_i * c_i, where c_i is an input scalar.
keras
edited Nov 13 '18 at 10:36
Abhishek Kumar
1,020216
asked Nov 13 '18 at 10:27
Filippo PorteraFilippo Portera
206
add a comment |
0
0
0
Todo :
- I would like to add a weight for each pattern loss in a given Keras loss function.
For example:
if the error on pattern i is l_i,
I would like to consider, instead, an error l_i * c_i, where c_i is an input scalar.
keras
edited Nov 13 '18 at 10:36
Abhishek Kumar
1,020216
asked Nov 13 '18 at 10:27
Filippo PorteraFilippo Portera
206
Todo :
- I would like to add a weight for each pattern loss in a given Keras loss function.
For example:
if the error on pattern i is l_i,
I would like to consider, instead, an error l_i * c_i, where c_i is an input scalar.
keras
keras
edited Nov 13 '18 at 10:36
Abhishek Kumar
1,020216
asked Nov 13 '18 at 10:27
Filippo PorteraFilippo Portera
206
edited Nov 13 '18 at 10:36
Abhishek Kumar
1,020216
asked Nov 13 '18 at 10:27
Filippo PorteraFilippo Portera
206
edited Nov 13 '18 at 10:36
Abhishek Kumar
1,020216
edited Nov 13 '18 at 10:36
Abhishek Kumar
1,020216
edited Nov 13 '18 at 10:36
Abhishek Kumar
1,020216
1,020216
asked Nov 13 '18 at 10:27
Filippo PorteraFilippo Portera
206
asked Nov 13 '18 at 10:27
Filippo PorteraFilippo Portera
206
asked Nov 13 '18 at 10:27
Filippo PorteraFilippo Portera
206
206
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
0
def customloss(y_true, y_pred):
c_i = ...
loss = ...(only use tensor operations on y_true and y_pred or use built in keras losses)
return c_i*loss
Now compile your model passing the loss function.
model.compile(loss = customloss)
answered Nov 13 '18 at 12:09
Mete Han KahramanMete Han Kahraman
40017
Sorry for my ignorance: are y_true and y_pred tensors (arrays in this case) covering all patterns? In other words, loss is the loss on a single pattern or the overall loss?
– Filippo Portera
Nov 13 '18 at 12:31
yes y_true and y_pred are tensors, when training your keras model will call this loss function to calculate the loss. Your loss function needs to return a single value tensor. So it needs to calculate overall loss.
– Mete Han Kahraman
Nov 13 '18 at 12:37
For my needs, I would consider inserting costs just inside the loss on a single pattern: loss_i * c_i
– Filippo Portera
Nov 13 '18 at 12:37
I mean: usually a loss is the sum of a function of y_true and y_pred for every pattern. I need to associate costs to a pattern level.
– Filippo Portera
Nov 13 '18 at 12:40
I think I've got your point!
– Filippo Portera
Nov 13 '18 at 12:41
|
show 16 more comments
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%2f53278915%2fkeras-how-to-add-weights-to-loss-evaluation%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
def customloss(y_true, y_pred):
c_i = ...
loss = ...(only use tensor operations on y_true and y_pred or use built in keras losses)
return c_i*loss
Now compile your model passing the loss function.
model.compile(loss = customloss)
answered Nov 13 '18 at 12:09
Mete Han KahramanMete Han Kahraman
40017
Sorry for my ignorance: are y_true and y_pred tensors (arrays in this case) covering all patterns? In other words, loss is the loss on a single pattern or the overall loss?
– Filippo Portera
Nov 13 '18 at 12:31
yes y_true and y_pred are tensors, when training your keras model will call this loss function to calculate the loss. Your loss function needs to return a single value tensor. So it needs to calculate overall loss.
– Mete Han Kahraman
Nov 13 '18 at 12:37
For my needs, I would consider inserting costs just inside the loss on a single pattern: loss_i * c_i
– Filippo Portera
Nov 13 '18 at 12:37
I mean: usually a loss is the sum of a function of y_true and y_pred for every pattern. I need to associate costs to a pattern level.
– Filippo Portera
Nov 13 '18 at 12:40
I think I've got your point!
– Filippo Portera
Nov 13 '18 at 12:41
|
show 16 more comments
0
def customloss(y_true, y_pred):
c_i = ...
loss = ...(only use tensor operations on y_true and y_pred or use built in keras losses)
return c_i*loss
Now compile your model passing the loss function.
model.compile(loss = customloss)
answered Nov 13 '18 at 12:09
Mete Han KahramanMete Han Kahraman
40017
Sorry for my ignorance: are y_true and y_pred tensors (arrays in this case) covering all patterns? In other words, loss is the loss on a single pattern or the overall loss?
– Filippo Portera
Nov 13 '18 at 12:31
yes y_true and y_pred are tensors, when training your keras model will call this loss function to calculate the loss. Your loss function needs to return a single value tensor. So it needs to calculate overall loss.
– Mete Han Kahraman
Nov 13 '18 at 12:37
For my needs, I would consider inserting costs just inside the loss on a single pattern: loss_i * c_i
– Filippo Portera
Nov 13 '18 at 12:37
I mean: usually a loss is the sum of a function of y_true and y_pred for every pattern. I need to associate costs to a pattern level.
– Filippo Portera
Nov 13 '18 at 12:40
I think I've got your point!
– Filippo Portera
Nov 13 '18 at 12:41
|
show 16 more comments
0
0
0
def customloss(y_true, y_pred):
c_i = ...
loss = ...(only use tensor operations on y_true and y_pred or use built in keras losses)
return c_i*loss
Now compile your model passing the loss function.
model.compile(loss = customloss)
answered Nov 13 '18 at 12:09
Mete Han KahramanMete Han Kahraman
40017
def customloss(y_true, y_pred):
c_i = ...
loss = ...(only use tensor operations on y_true and y_pred or use built in keras losses)
return c_i*loss
Now compile your model passing the loss function.
model.compile(loss = customloss)
answered Nov 13 '18 at 12:09
Mete Han KahramanMete Han Kahraman
40017
answered Nov 13 '18 at 12:09
Mete Han KahramanMete Han Kahraman
40017
answered Nov 13 '18 at 12:09
Mete Han KahramanMete Han Kahraman
40017
answered Nov 13 '18 at 12:09
Mete Han KahramanMete Han Kahraman
40017
40017
Sorry for my ignorance: are y_true and y_pred tensors (arrays in this case) covering all patterns? In other words, loss is the loss on a single pattern or the overall loss?
– Filippo Portera
Nov 13 '18 at 12:31
yes y_true and y_pred are tensors, when training your keras model will call this loss function to calculate the loss. Your loss function needs to return a single value tensor. So it needs to calculate overall loss.
– Mete Han Kahraman
Nov 13 '18 at 12:37
For my needs, I would consider inserting costs just inside the loss on a single pattern: loss_i * c_i
– Filippo Portera
Nov 13 '18 at 12:37
I mean: usually a loss is the sum of a function of y_true and y_pred for every pattern. I need to associate costs to a pattern level.
– Filippo Portera
Nov 13 '18 at 12:40
I think I've got your point!
– Filippo Portera
Nov 13 '18 at 12:41
|
show 16 more comments
Sorry for my ignorance: are y_true and y_pred tensors (arrays in this case) covering all patterns? In other words, loss is the loss on a single pattern or the overall loss?
– Filippo Portera
Nov 13 '18 at 12:31
yes y_true and y_pred are tensors, when training your keras model will call this loss function to calculate the loss. Your loss function needs to return a single value tensor. So it needs to calculate overall loss.
– Mete Han Kahraman
Nov 13 '18 at 12:37
For my needs, I would consider inserting costs just inside the loss on a single pattern: loss_i * c_i
– Filippo Portera
Nov 13 '18 at 12:37
I mean: usually a loss is the sum of a function of y_true and y_pred for every pattern. I need to associate costs to a pattern level.
– Filippo Portera
Nov 13 '18 at 12:40
I think I've got your point!
– Filippo Portera
Nov 13 '18 at 12:41
Sorry for my ignorance: are y_true and y_pred tensors (arrays in this case) covering all patterns? In other words, loss is the loss on a single pattern or the overall loss?
– Filippo Portera
Nov 13 '18 at 12:31
Sorry for my ignorance: are y_true and y_pred tensors (arrays in this case) covering all patterns? In other words, loss is the loss on a single pattern or the overall loss?
– Filippo Portera
Nov 13 '18 at 12:31
yes y_true and y_pred are tensors, when training your keras model will call this loss function to calculate the loss. Your loss function needs to return a single value tensor. So it needs to calculate overall loss.
– Mete Han Kahraman
Nov 13 '18 at 12:37
yes y_true and y_pred are tensors, when training your keras model will call this loss function to calculate the loss. Your loss function needs to return a single value tensor. So it needs to calculate overall loss.
– Mete Han Kahraman
Nov 13 '18 at 12:37
For my needs, I would consider inserting costs just inside the loss on a single pattern: loss_i * c_i
– Filippo Portera
Nov 13 '18 at 12:37
For my needs, I would consider inserting costs just inside the loss on a single pattern: loss_i * c_i
– Filippo Portera
Nov 13 '18 at 12:37
I mean: usually a loss is the sum of a function of y_true and y_pred for every pattern. I need to associate costs to a pattern level.
– Filippo Portera
Nov 13 '18 at 12:40
I mean: usually a loss is the sum of a function of y_true and y_pred for every pattern. I need to associate costs to a pattern level.
– Filippo Portera
Nov 13 '18 at 12:40
I think I've got your point!
– Filippo Portera
Nov 13 '18 at 12:41
I think I've got your point!
– Filippo Portera
Nov 13 '18 at 12:41
|
show 16 more comments
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%2f53278915%2fkeras-how-to-add-weights-to-loss-evaluation%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
