minstar

> @minstar Did you find the reason for this? I just used HuggingFace code (run_ner.py) to reproduce results. And as I think, the pad position only needs to be considered...

Hi @leonweber, thank you for your prompt response!

Hi @leonweber, sorry for reopening the issue. I want to reproduce an in-domain evaluation (Table 2) of Hunflair paper. Which model do I have to use it?

Hi @liwenqingi If you have data of a new type, then you could modify and train through the modeling.py file to set up a separate classifier.

I prefer to choose the latter case, this is because you have to find optimal training settings which could be time-consuming and labor-intensive things. May I ask the reason that...

Then, how about adding your new entity classifying system through socket communication as we did? In bern2.py, line 361-363, we separately get the results of tmvar, gnormplus, and our multi-ner...