So I'm training a sequence-labeling model with a BiLSTM-CRF architecture, and I am getting negative values on Viterbi Loss. Is this possible?

I'm using the following formula in my code, as specified from this tutorial: $$ \begin{equation} L = \log({\textstyle \sum_{t \in T}} e^{\text{score(t)}})-\text{score}(t_G) \end{equation} $$ I used this as is, without getting its absolute value.


Your Answer

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

Browse other questions tagged or ask your own question.