urchade

/

gliner_multi_pii-v1

like 57

I am looking to fine tune the gliner_multi_pii_v1 on a NER dataset containing 50K samples with 16 different entities each occurring ~ 17k times across the whole dataset. This dataset is balanced for 16 diff entities. I want to know the range in terms of epoch or steps for the model to perform decently.

For how many steps/epochs the gliner_multi_pii_v1 was trained?

I am also looking to use the generalised capability of the model for my non-primary use-cases for which I am not doing the training and will use the model off-the-shelf.