Open ChengchengDu opened 2 years ago
log σ(v0si−1> vsi ) + log σ(v0si+1
vsi ) + Es∼P(s) log σ(−v0s> vs)] (3) where vs and v0s are two different representations of sentence s via two differently parameterized BERT encoders;
in the objective of finetune bert, why use two different ly parameterized bert encoder?
Hi, I cound't find the code of how to finetune the bert. Could you give me some guidance? Thank you very much