Evaluation Metrics.

[RishitJoshi97]

First of all thank you for providing model files,due to limited resources i was not able to train a good model.I wanted to evaluate the predictions as done in original work here(https://github.com/smitkiri/ehr-relation-extraction/tree/master/biobert_re).It seems it is done using Track2-evaluate-ver4.py file.I tried to do it but it requires predictions in same format as in .ann files. If you have produce the results using Track2-evaluate-ver4.py file can you enlighten me on that please.

[vedants03]

I had the same problem too, however I was unable to figure out the solution. I ran the evaluation on test and evaluation dataset and not on gold standard dataset. You can find more information regarding the issue here.