Few Shot Text Classification

I provided some minor clean-ups/tests for a library that demonstrates zero-shot and few-shot classification of documents.