how can you train a model

[Cristian Aguirre]

I need to create a model, because the result is always varied, initially line 1 has the text: "Surnames...", line 2: "Names..."; but when passing the recognition in another photo this changes line 2, now it has the name. And line 0, the last name. And so it always varies, it would be useful to generate a model and put it to use in recognition. Right now I am working with version 2.0.0-alpha.13 of javascript, this is because I return the progress and text as part of the same object/dictionary.
My question would be how to create and implement a model in .JS of tesseract

In advance, thanks for any help, I've been reviewing the documentation but with so many tabs I lost it, although that was for a current version and not in .JS, although the version doesn't matter much, it does affect me that it's not in .JS It is a project that closes in a few days