Commit Graph

3 Commits (709dfd7e81261ac23658cffbee36afbcdc765fea)

Author SHA1 Message Date
qianlong cae77c0c22 BasicTokenizer not case fold on preserverd words
6 years ago
qianlong 980ddd32a2 change output of WordpieceTokenizer and BertTokenizer to 1-D string tensors
6 years ago
qianlong 4f16f036be Add WhitespaceTokenizer and UnicodeScriptTokenizer for nlp
6 years ago