Add transformers to requirements

[ericschreiber]

No description provided.

[tugot17]

where do we use this?

[iPieter]

I am having the same issue, just replicating the readme's instruction.

In TextSampler, you load a tokenizer from the transformer library:

class TextSampler:
    ... 
    def __init__(self, 
                 tokenizer_name: str,
                 dataset_loader: DatasetLoader):
        ...
        self.tokenizer = AutoTokenizer.from_pretrained(tokenizer_name)