1.4.2

Bug fixes

1.4.1

Expose transformer hidden_size parameter

1.4

Wrapper around Tokenizer output. Lot of bug fixes.

1.3.2

Update Tokenizer parameters, fixed batch bug.

1.3.1

Fix import bug

1.3

Added SpaCy tokenizer. Set use_spacy=True to use it. By default, uses a multilingual model. It can be changed during Tokenizer init with language parameter.

1.1

Add split_on_space flag to tokenizer, update README.md, fix some bugs.

1.0

First release.