Hello,
I am using the PyTorch
version of Hugging Face library’s BERT
model and I want to continue the pre-training of the model it in a domain specific dataset, before fine-tune it.
What is the best way to do it?
General tips about pre-training are much appreciated.