class BartTokenizer(vocab_file, merges_file, errors='replace', bos_token='<s>', eos_token='</s>', cls_token='<s>', sep_token='</s>', unk_token='<unk>', pad_token='<pad>', mask_token='<mask>', **kwargs)[source]#

Bases: PretrainedTokenizer

Construct a BART tokenizer based on byte-level Byte-Pair-Encoding.

This tokenizer inherits from GPTTokenizer. For more information regarding those methods, please refer to this superclass.

  • vocab_file (str) – Path to the vocabulary file. The vocab file contains a mapping from vocabulary strings to indices.

  • merges_file (str) – Path to the merge file. The merge file is used to split the input sentence into “subword” units. The vocab file is then used to encode those units as intices.

  • errors (str) – Paradigm to follow when decoding bytes to UTF-8. Defaults to 'replace'.

  • max_len (int, optional) – The maximum value of the input sequence length. Defaults to None.

  • bos_token (str, optional) – The beginning of sequence token that was used during pretraining. Can be used a sequence classifier token. Defaults to "<s>".

  • eos_token (str, optional) – A special token representing the end of a sequence that was used during pretraining. Defaults to "</s>".

  • cls_token (str, optional) – A special token used for sequence classification. It is the last token of the sequence when built with special tokens. Defaults to "<s>".

  • sep_token (str, optional) – A special token separating two different sentences in the same input. Defaults to "</s>".

  • unk_token (str, optional) – A special token representing the unknown (out-of-vocabulary) token. An unknown token is set to be unk_token inorder to be converted to an ID. Defaults to "<unk>".

  • pad_token (str, optional) – A special token used to make arrays of tokens the same size for batching purposes. Defaults to "<pad>".

  • mask_token (str, optional) – A special token representing a masked token. This is the token used in the masked language modeling task which the model tries to predict the original unmasked ones. Defaults to "<mask>".


from paddlenlp.transformers import BartTokenizer

tokenizer = BartTokenizer.from_pretrained('bart-base')
print(tokenizer('He was a puppeteer'))

{'input_ids': [0, 894, 21, 10, 32986, 9306, 254, 2],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1]}
build_inputs_with_special_tokens(token_ids_0, token_ids_1=None)[source]#

Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and adding special tokens.

get_special_tokens_mask(token_ids_0, token_ids_1=None, already_has_special_tokens=False)[source]#

Retrieves sequence ids from a token list that has no special tokens added. This method is called when adding special tokens using the tokenizer encode methods.

create_token_type_ids_from_sequences(token_ids_0, token_ids_1=None)[source]#

Create a mask from the two sequences passed to be used in a sequence-pair classification task.


Returns the vocabulary as a dictionary of token to index.

tokenizer.get_vocab()[token] is equivalent to tokenizer.convert_tokens_to_ids(token) when token is in the vocab.


The vocabulary.

Return type:

Dict[str, int]

property vocab_size#

Returns the size of vocabulary.


The sum of size of vocabulary and the size of speical tokens.

Return type:



Converts a single index or a sequence of indices to texts.


ids (int|List[int]) – The token id (or token ids) to be converted to text.


The decoded text.

Return type:



from paddlenlp.transformers import GPTTokenizer
tokenizer = GPTTokenizer.from_pretrained('gpt2-medium-en')
print(tokenizer.convert_ids_to_string(tokenizer.convert_ids_to_string([14618, 284, 779, 350, 37382, 47, 37382, 290, 350, 37382, 45, 19930]))
# 'Welcome to use PaddlePaddle and PaddleNLP'

Saves SentencePiece file (ends with ‘.spm’) under save_directory.


save_directory (str) – Directory to save files into.


Converts a sequence of tokens (string) in a single string.

build_offset_mapping_with_special_tokens(offset_mapping_0, offset_mapping_1=None)[source]#

Build offset map from a pair of offset map by concatenating and adding offsets of special tokens.

A BERT offset_mapping has the following format:

  • single sequence: (0,0) X (0,0)

  • pair of sequences: (0,0) A (0,0) B (0,0)

  • offset_mapping_ids_0 (List[tuple]) – List of wordpiece offsets to which the special tokens will be added.

  • offset_mapping_ids_1 (List[tuple], optional) – Optional second list of wordpiece offsets for offset mapping pairs. Defaults to None.


A list of wordpiece offsets with the appropriate offsets of special tokens.

Return type: