How to use the ELMo/BERT model to do fine tune
Anikacyp opened this issue · 1 comments
Anikacyp commented
It seems ELMo/BERT has the same functions as glove, Can you tell me how to use the hidden size of BERT to do fine-tune. Say train the BERT parameters at the same time.
allanj commented
At the moment, we are not supporting this feature yet. I plan to support this soon but the next ACL deadline is approaching. But what I would suggest is to incorporate the transformer package by hugging face. (https://github.com/huggingface/transformers)