text classification with large amount of data

Question

I am working on a text classification project. First, i collected more than 3 GB dataset(text). these data contains more than 600 labels. my question is : i think working on all