When fine-tuning a BERT model using the medkit library I had to compare it to a fine-tuning process made directly with the hugging face library and I stood upon the fact that I couldn't replicate exactly what was done with the HF library in medkit.
Options such as weight decay and or early stopping could not be implemented.
I don't know if it would be possible to have more access to the hugging face training library directly trough the medkit-lib but it would help having more options when training models.