You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
SSDKL is an excellent for semi-supervised learning. I want to apply this method to solve my problem. However, I have a question about the unlabelled data.
In SSDKL, the labelled data was divided into training, validating, testing and unlabelled data. In my understanding, the label (target values) of the unlabelled section should not be fed into the training model. However, from the functions of "def _setup_nns(self):" and "def _compute_embeddings(self, sess):", the label "self.y_unlabeled" is also an input for NN training. Since my data contains real unlabelled data without any labells, if so, it seems that I cannot adapt my data to the SSDKL method directly.
Did I understand right?
Thanks for your help~
Best regards,
Jappy
The text was updated successfully, but these errors were encountered:
Hi,
The method does not use labels on the unlabeled data for training. We may
be using them just for logging some metrics, but they are not going into
the loss. You should be able to use it for your purposes!
On Fri, Dec 17, 2021 at 8:47 PM Jappy ***@***.***> wrote:
Hi,
SSDKL is an excellent for semi-supervised learning. I want to apply this
method to solve my problem. However, I have a question about the unlabelled
data.
In SSDKL, the labelled data was divided into training, validating, testing
and unlabelled data. In my understanding, the label (target values) of the
unlabelled section should not be fed into the training model. However, from
the functions of "def _setup_nns(self):" and "def _compute_embeddings(self,
sess):", the label "self.y_unlabeled" is also an input for NN training.
Since my data contains real unlabelled data without any labells, if so, it
seems that I cannot adapt my data to the SSDKL method directly.
Did I understand right?
Thanks for your help~
Best regards,
Jappy
—
Reply to this email directly, view it on GitHub
<#5>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AB4J2SSD2FPWZGO6DRIRSITURPY3DANCNFSM5KKDWWHA>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
Hi,
SSDKL is an excellent for semi-supervised learning. I want to apply this method to solve my problem. However, I have a question about the unlabelled data.
In SSDKL, the labelled data was divided into training, validating, testing and unlabelled data. In my understanding, the label (target values) of the unlabelled section should not be fed into the training model. However, from the functions of "def _setup_nns(self):" and "def _compute_embeddings(self, sess):", the label "self.y_unlabeled" is also an input for NN training. Since my data contains real unlabelled data without any labells, if so, it seems that I cannot adapt my data to the SSDKL method directly.
Did I understand right?
Thanks for your help~
Best regards,
Jappy
The text was updated successfully, but these errors were encountered: