0
$\begingroup$

I followed the steps to feed in both dev, test, train.tsv to the model, trained it, then tried to classify test data, and I only have 1 feature, and the classification is binary, 1 or 0. I assumed my test_results.tsv would have just 1 column, representing the estimated classification. Any advice? Here is a picture of my output. I thought it would be a column of 1's and 0's showing me its estimated classification.

enter image description here

$\endgroup$
1
  • $\begingroup$ Hi @AJT, welcome to the site. We need more information to help. What implementation of BERT are you using? Where did you take your classification code from? $\endgroup$
    – noe
    Nov 22 '20 at 9:03
0
$\begingroup$

If you look at your output you will see that the two values in every row add up to 1. This means that the first value is the probability of the output having value 0 in your classification, and the second value the probability of the out having value 1. If you want to only have the 0/1 classification instead of the probabilities you can use a simply threshold for the probabilities (i.e. if value < threshold then 0, else 1). The simplest way to set the threshold could be 0.5, but you could try changing threshold and see how the performance on your test set changes.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.