TransformerWordEmbeddings wrong arguments in README: batch_size not working

[ugm2]

Describe the bug
I was following TRANSFORMER_EMBEDDINGS.md tutorial and tried assigning a batch_size for multiple sentences inference at once, and I got an error saying that batch_size is an unexpected keyword argument.

In the Arguments Section it says that I can use batch_size, but I was unable to.

It seems in Flair==0.7 works

To Reproduce

from flair.embeddings import TransformerWordEmbeddings
model = TransformerWordEmbeddings('bert-base-uncased', batch_size=2)

Traceback:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/unai/miniconda3/envs/li_encoding/lib/python3.7/site-packages/flair/embeddings/token.py", line 821, in __init__
    self.model = AutoModel.from_pretrained(model, config=config, **kwargs)
  File "/home/unai/miniconda3/envs/li_encoding/lib/python3.7/site-packages/transformers/models/auto/modeling_auto.py", line 815, in from_pretrained
    pretrained_model_name_or_path, *model_args, config=config, **kwargs
  File "/home/unai/miniconda3/envs/li_encoding/lib/python3.7/site-packages/transformers/modeling_utils.py", line 1058, in from_pretrained
    model = cls(config, *model_args, **model_kwargs)
TypeError: __init__() got an unexpected keyword argument 'batch_size'

Expected behavior
I would expect a normal initialization of model, and then being able to embed sentences in chunks

Environment:

• OS: Linux (Ubuntu 18.04.5 LTS)
• Flair Version: 0.8

[alanakbik]

Hello @ugm2 we removed the batch_size parameter in Flair 0.8. I'll update the documentation!

[alanakbik]

Updated the docs, thanks for spotting this!