| name: bpe_tokenizer | |
| config_type: preprocessor | |
| pretrained_path: hezarai/roberta-base-fa | |
| max_length: 512 | |
| truncation_strategy: longest_first | |
| truncation_direction: right | |
| stride: 0 | |
| padding_strategy: longest | |
| padding_direction: right | |
| pad_to_multiple_of: 0 | |
| pad_token_id: 0 | |
| pad_token: <pad> | |
| pad_token_type_id: 0 | |
| unk_token: <unk> | |
| special_tokens: | |
| - <s> | |
| - <pad> | |
| - </s> | |
| - <unk> | |
| - <mask> | |
| - <|endoftext|> | |
| - <|startoftext|> | |
| - <nl> | |
| - <hs> | |
| - <sep> | |
| - <cls> | |
| continuing_subword_prefix: '' | |
| end_of_word_suffix: '' | |
| fuse_unk: false | |
| train_config: | |
| name: bpe_tokenizer | |
| config_type: preprocessor | |
| vocab_size: 30000 | |
| min_frequency: 2 | |
| limit_alphabet: 1000 | |
| show_progress: true | |