Text Classifcation with Short Text

I have 3 dataframes( train, val and test), each containing a field “class” and a field “text”, and I want to create a neural network to predict the class from a given text.

The text is domain specific and short, so I can not create a Language Model to predict the next word as a first step.

I want to use LanguageModelData.from_dataframes() to create a model, and then train it to predict the class.

How can I tell the LanguageModelData where the classes are stored in the dataframe?

Thanks a lot,

I forgot to mention that we need to predict 50 classes based on the text.
Kind regards