kentonl/e2e-coref

can transfer learning work with the given pretrained model ?

Opened this issue · 0 comments

Hi,

Thanks for sharing good work. it's helping me to solve one my work called "clustering/gathering sentences w.r.t entity".
But I'm not able to use it for my domain(finance) data. so i am thinking of creating small amount (max ~500 samples/examples)of training data for same domain and then use transfer learning to learn. is it possible ? will it learn only with 500 examples or some minimum number of samples required ?

Thanks in advance.