Good morning engaged community. I’m having difficulty with marckupLM, I would like to know if there is a way for the MarkupLMProcessor to consider information that was truncated for the data itself. My question comes from the need to have larger data than I had to truncate, but I need the model to consider the information that was truncated for the classification of tokens.

I was hoping there would be some parameter that facilitated data with different sizes, but during my research I haven’t found it yet. I would like to know if I really have to fill it out by hand, or if there is some way.