fixed a typo in README.md

made a minor cosmetic update to the "Dataset" section

updated README.md

pushing tokenizer

epoch: 2, f1: 0.9354945968412303, accuracy: 0.9858421145581916

epoch: 1, f1: 0.9325132978723404, accuracy: 0.9857685288750221

epoch: 0, f1: 0.9163963064636886, accuracy: 0.9790133631600636

initial commit