Closed luckysusanto closed 2 months ago
The paper mentioned the use of BERT-score. Does it use the base pre-trained BERT? Or is it further fine-tuned?
It is a base pretrained BERT. You can check it in https://arxiv.org/pdf/1904.09675 and https://github.com/Tiiiger/bert_score
Thanks! Solved.
The paper mentioned the use of BERT-score. Does it use the base pre-trained BERT? Or is it further fine-tuned?