Skip to content

1.2.0: Bert Pipe
Compare
Choose a tag to compare
@SoluMilken SoluMilken released this 24 Jan 09:15
· 149 commits to master since this release
1.2.0
b461d0c
  1. Add building blocks for BERT tokenizer construction, including AddWhitespaceAroundCJK, AddWhitespaceAroundPunctuation, MergeWhiteSpaceCharacters, StripWhiteSpaceCharacters,
    StripAccentToken, WhiteSpaceTokenizer and SpanSubwords.
  2. Bert pipes are created in uttut/pipeline/bert/.
Assets 2
Loading