Add use_tokenizer to JsonlDataset

[david-waterworth]

This PR adds use_tokenizer argument to the SequenceTagger JSONL datasets. This allows use of non-default tokenization with taggers trained using JSONL datasets.

Closes #3476

[alanakbik]

Hello @david-waterworth thanks for contributing! One of our checks is failing since the code was not formatted with black.

(See https://github.com/flairNLP/flair/blob/master/CONTRIBUTING.md#code-formatting)

To fix it, could you run black to auto-format your code and push again?

[david-waterworth]

@alanakbik I've fixed the formatting

[alanakbik]

@david-waterworth thanks for adding this!