train my IOB format using spacy

[myeghaneh]

I have data in this format, (IOB format on token level)

[On, IN, O], [the, DT, O], [other, JJ, O], [hand, NN, O], [,, ,, O], [the, DT, B-P], [significance, NN, I-P], [of, IN, I-P], [competition, NN, I-P], [is, VBZ, I-P], [that, IN, I-P], [how, WRB, I-P], [to, TO, I-P], [become, VB, I-P], [more, JJR, I-P], [excellence, NN, I-P], [to, TO, I-P], [gain, VB, I-P], [the, DT, I-P], [victory, NN, I-P], [., ., O], [Hence, RB, O], [it, PRP, O], [is, VBZ, O], [always, RB, O], [said, VBN, O], [that, IN, O], [competition, NN, B-C], [makes, VBZ, I-C], [the, DT, I-C], [society, NN, I-C], [more, RBR, I-C], [effective, JJ, I-C], [., ., O],

how can I train a classifer which work like NER by spacy? (My case is not NER , but the format is close)

should I convert my data to spacy format?

[aakashb95]

Any luck with this?

[ljvmiranda921]

Hi @myeghaneh , if you have data in the IOB format, perhaps you can check out the spaCy convert command? You can refer to these examples for more info. It might even make sense to save them as an .iob file first before running the command.

Once you have the .spacy files, you can now start training.