tflite_support.metadata_writers.metadata_info.RegexTokenizerMd

A container for the Regex tokenizer 1 metadata information.

tflite_support.metadata_writers.metadata_info.RegexTokenizerMd(
    delim_regex_pattern: str, vocab_file_path: str
)

delim_regex_pattern the regular expression to segment strings and create tokens.
vocab_file_path path to the vocabulary file.

Methods

create_metadata

View source

create_metadata() -> tflite_support.metadata_schema_py_generated.ProcessUnitT

Creates the Bert tokenizer metadata based on the information.

Returns
A Flatbuffers Python object of the Bert tokenizer metadata.