Fix data error that prevents Deep Learning Model to train properly

I have a project in which I need some support regarding a Visual Question Generation model in Radiology. Unfortunately, I am having an error while the model is training that from what I have researched online is due to some empty lines in one of the files that is used to train the model. I tried a few solutions but unfortunately with no luck which I can go into detail if we have a call. If everything goes well with this version of the model, I would like for us to continue our cooperation because even after this error is fixed and the model runs, I will need support to switch the LSTM part of the model with a Transformer architecture as part of my next task regarding this project. The deep learning model implementation is based on the following github repo: https://github.com/sarrouti/VQGRaD.git