Question about parameters for further training of a preexisiting model?

[mcriggs]

Hi!

First off, thank you in advance for any advice you can offer.

I would like to further train a preexisting model for a specific domain and I hope to be able to do so using your train.py script.

I have already prepared my data in accordance with your instructions. That is, I have already processed my data according to all the steps outlined in "Preparing the data" (https://github.com/facebookresearch/XLM#1-preparing-the-data).

In order to further train a preexisting model on this data, can I simply run your train.py script with the parameters listed in your "Train the BERT model" (https://github.com/facebookresearch/XLM#2-train-the-bert-model) with the addition of a parameter that pertains to the preexisting model? For example, could I simply add an additional parameter like "--model_path preexisting_model.pth" ?

I apologize if the answer to this question is obvious!

Many thanks. :)