We introduce a two-phase fine-tuning 012 strategy: first pre-training on the Financial PhraseBank dataset, then adapting to the Twitter Financial News Sentiment dataset. - View it on GitHub