How to "re-train" BioBert on a custom medical corpus?

[peterphancong]

Thank you for you wonderful BioBert pretrained model. As a part of my work, I would like to have BioBert model trained in my medical corpus (without labels) before use it for further text embedding. However, all the source code I found mostly require labelled data (finetuning not further pretraining). Could you please introduce a sample for that?

Thank you very much