This is not a problem with calc() method. Calc method is to compare the
labelling done by the model with the tagged data and for this the
dataSequence(input sequence) must be the same. tokenAuto.length !=
tokenMan.length clearly means that the input sequence is not the same (the
'length' is the number of tokens in the dataSequence).
Also, could you please elaborate on the following statement. The length
parameter as mentioned above is the number of tokens in the dataSequence and
how you are relating with the numbers of labels inferred is not clear.
This case happens when for a particular sequence in
the tag file, there are more labels than inferred by
applying the model.
Get latest updates about Open Source Projects, Conferences and News.