arxyzan's picture
Update preprocessor/tokenizer_config.yaml
b584b5e
raw
history blame
265 Bytes
name: wordpiece_tokenizer
config_type: preprocessor
model: wordpiece
max_length: 512
truncation_strategy: longest_first
truncation_direction: right
stride: 0
padding_strategy: longest
padding_direction: right
pad_token_id: 0
pad_token: '[PAD]'
pad_token_type_id: 0