o
ei @ sv d Z ddlZddlmZmZmZmZ ddlmZ ddl m
Z
ddl m Z e
eZdd d
ZG d d d e
Zd gZdS )
z Tokenization class for model T5. N) Tokenizerdecoderspre_tokenizers
processors)Unigram )TokenizersBackend)loggingz spiece.modelztokenizer.json)
vocab_filetokenizer_filec sh e Zd ZdZeZddgZeZ dd e e
e e e f B dB f fd
d
Z
d d
Zdd Z ZS ) T5Tokenizera
Construct a T5 tokenizer (backed by HuggingFace's *tokenizers* library). Based on
[Unigram](https://huggingface.co/docs/tokenizers/python/latest/components.html?highlight=unigram#models).
This tokenizer inherits from [`TokenizersBackend`] which contains most of the main methods. Users should
refer to this superclass for more information regarding those methods.
Args:
vocab_file (`str`, *optional*):
[SentencePiece](https://github.com/google/sentencepiece) file (generally has a *.spm* extension) that
contains the vocabulary necessary to instantiate a tokenizer.
eos_token (`str`, *optional*, defaults to `""`):
The end of sequence token.
When building a sequence using special tokens, this is not the token that is used for the end of sequence.
The token used is the `sep_token`.
unk_token (`str`, *optional*, defaults to `""`):
The unknown token. A token that is not in the vocabulary cannot be converted to an ID and is set to be this
token instead.
pad_token (`str`, *optional*, defaults to `""`):
The token used for padding, for example when batching sequences of different lengths.
extra_ids (`int`, *optional*, defaults to 100):
Add a number of extra ids added to the vocabulary for use as sentinels. These tokens are accessible as
"" where "{%d}" is a number between 0 and extra_ids-1. These tokens can be retrieved by
calling get_sentinel_tokens method and token ids can be by calling get_sentinel_token_ids method
additional_special_tokens (`list[str]`, *optional*):
Additional special tokens used by the tokenizer.
vocab (`str`, `dict` or `list`, *optional*):
Custom vocabulary dict. If not provided, a minimal vocabulary is created using the special tokens.
input_idsattention_maskNd vocabc
s || _ |d ur6dd |D }t|dk r |dd t|D 7 }n!|dkr5|t|kr5td| d| dn d d t|D }|}|d urI|| _n*t|d
ft|d
ft|d
fd g| _t|d d d D ]} | jd
| dd
f qdtt| jddd| _ d | j _
t t
t jddddg| j _tjdddd| j _t jd|||||d| tjddgg dd| jfgd| j _d S ) Nc S s g | ]
}d t |v r|qS )
U s z(T5Tokenizer.__init__.. c S g | ]}d | dqS r >r r ir r r r W r zBoth extra_ids (z!) and additional_special_tokens (zk) are provided to T5Tokenizer. In this case the additional_special_tokens must include the extra_ids tokensc S r r r r r r r r _ r! g ) ▁g r r F)unk_id
byte_fallbackr" alwaysT) replacementprepend_schemesplit) eos_token unk_token pad_token extra_idsadditional_special_tokens$Ar )r0 r z$Br )singlepairspecial_tokensr )
_extra_idslenrange
ValueError
_vocab_scoresr appendr r
_tokenizer
normalizerr SequenceWhitespaceSplit Metaspace
pre_tokenizerr decodersuper__init__r TemplateProcessing eos_token_idpost_processor)
selfr r+ r, r- r. r/ kwargs extra_tokensr __class__r r rB G sd
zT5Tokenizer.__init__c C s t ttdd | jS )zQGet the list of sentinel tokens (extra_id tokens) from additional_special_tokens.c S s t td| d uS )Nz)boolresearch)r r r r s z1T5Tokenizer.get_sentinel_tokens..)listsetfilterr/ rF r r r get_sentinel_tokens s z T5Tokenizer.get_sentinel_tokensc s fdd D S )z&Get the token IDs for sentinel tokens.c s g | ]} |qS r )convert_tokens_to_ids)r tokenrR r r r s z6T5Tokenizer.get_sentinel_token_ids..)rS rR r rR r get_sentinel_token_ids s z"T5Tokenizer.get_sentinel_token_ids)Nr r r r N)__name__
__module__ __qualname____doc__VOCAB_FILES_NAMESvocab_files_namesmodel_input_namesr modelr rO tuplefloatrB rS rV
__classcell__r r rI r r s $Lr )rZ rL
tokenizersr r r r tokenizers.modelsr tokenization_utils_tokenizersr utilsr
get_loggerrW loggerr[ r __all__r r r r s