How to fine tune model the same way as the pre-trained model

[AlejandroRigau]

I want to fine tune a model using the same learning rate, optimizer, etc. as the original pre-trained model used. How can I do this using the train.py script?