# loading the tokenizer and vocab processors
processor = get_roberta_processor(tokenizer=fastai_tokenizer, vocab=fastai_roberta_vocab)

# creating our databunch
data = ItemLists(".", RobertaTextList.from_df(train, ".", cols=feat_cols, processor=processor),
                      RobertaTextList.from_df(val, ".", cols=feat_cols, processor=processor)
                ) \
       .label_from_df(cols=label_cols, label_cls=CategoryList) \
       .add_test(RobertaTextList.from_df(test, ".", cols=feat_cols, processor=processor)) \
       .databunch(bs=4,pad_first=False)