Add pad token

Update README.md

Add tokenizer.json and simplify config.json

Create config.json

Make file name lowercase

Pack all tensors into one dataset

Create tokenizer.json

Remove `datasets` group

initial commit