U
4Af-' @ sZ d Z ddlZddlmZmZmZ ddlmZmZ ddl m
Z
e
e Z
G dd deZdS ) z"Tokenization class for model ByT5. N)ListOptionalTuple )
AddedTokenPreTrainedTokenizer)loggingc s e Zd ZdZddgZd(dd fd
d
Zed d
Zdd Zd)e e
e e e
e e e
d fdd
Z
e e
e e
dddZd*e e
e e e
e e
dddZd+e e
e e e
e e
dddZee e d d d Zd d Zd!d" Zd#d$ Zd,ee e ee d%d&d'Z ZS )-
ByT5Tokenizera
Construct a ByT5 tokenizer. ByT5 simply uses raw bytes utf-8 encoding.
This tokenizer inherits from [`PreTrainedTokenizer`] which contains most of the main methods. Users should refer to
this superclass for more information regarding those methods.
Args:
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 125):
Add a number of extra ids added to the end of the vocabulary for use as sentinels. These tokens are
accessible as "" where "{%d}" is a number between 0 and extra_ids-1. Extra tokens are
indexed from the end of the vocabulary up to beginning ("" is the last token in the vocabulary
like in ByT5 preprocessing see
[here](https://github.com/google-research/text-to-text-transfer-transformer/blob/9fd7b14a769417be33bc6c850f9598764913c833/t5/data/preprocessors.py#L2117)).
additional_special_tokens (`List[str]`, *optional*):
Additional special tokens used by the tokenizer.
Z input_idsZattention_mask} N)returnc s
|dkr$|d kr$dd t |D }nP|dkrt|d k rtt|dkrttttdd |}||krttd| d| dt|trt|d d d
n|}t|trt|d d d
n|}t|trt|d d d
n|}|||d | _t| j| _ d | _
t j f |||d|d
| d S )Nr c S s g | ]}d | dqS )z
.0ir r N/tmp/pip-unpacked-wheel-zw5xktn0/transformers/models/byt5/tokenization_byt5.py
I s z*ByT5Tokenizer.__init__..c S s t dt| kS )NZextra_id)boolstr)xr r r L z(ByT5Tokenizer.__init__..zBoth extra_ids (z!) and additional_special_tokens (zm) are provided to ByT5Tokenizer. In this case the additional_special_tokens must include the extra_ids tokensT)lstriprstrip)r ) eos_token unk_token pad_token extra_idsadditional_special_tokens)
rangelensetfilter
ValueError
isinstancer r Z_added_tokens_decoderoffset_utf_vocab_sizesuper__init__)selfr r! r" r# r$ kwargsZ extra_tokens __class__r r r. > s.
zByT5Tokenizer.__init__c C s | j S )N)r, r/ r r r
vocab_sizee s zByT5Tokenizer.vocab_sizec s. fddt j j D }| j |S )Nc s i | ]} ||qS r )Zconvert_ids_to_tokensr r3 r r
j s z+ByT5Tokenizer.get_vocab..)r% r4 r+ updateadded_tokens_encoder)r/ Zvocabr r3 r get_vocabi s zByT5Tokenizer.get_vocabF) token_ids_0 token_ids_1already_has_special_tokensr c sZ |rt j||ddS |dkr2dgt| dg S dgt| dg dgt| dg S )a
Retrieve sequence ids from a token list that has no special tokens added. This method is called when adding
special tokens using the tokenizer `prepare_for_model` method.
Args:
token_ids_0 (`List[int]`):
List of IDs.
token_ids_1 (`List[int]`, *optional*):
Optional second list of IDs for sequence pairs.
already_has_special_tokens (`bool`, *optional*, defaults to `False`):
Whether or not the token list is already formatted with special tokens for the model.
Returns:
`List[int]`: A list of integers in the range [0, 1]: 1 for a special token, 0 for a sequence token.
T)r9 r: r; Nr r )r- get_special_tokens_maskr& )r/ r9 r: r; r1 r r r< n s z%ByT5Tokenizer.get_special_tokens_mask) token_idsr c C sB t |dkr2|d | jkr2td| j d |S || jg S dS )z.Do not add eos again if user already added it.r zThis sequence already has zQ. In future versions this behavior may lead to duplicated eos tokens being added.N)r& eos_token_idwarningswarnr )r/ r= r r r _add_eos_if_not_present s z%ByT5Tokenizer._add_eos_if_not_present)r9 r: r c C s< | j g}|dkr"t|| dg S t|| | | dg S )a
Create a mask from the two sequences passed to be used in a sequence-pair classification task. ByT5 does not
make use of token type ids, therefore a list of zeros is returned.
Args:
token_ids_0 (`List[int]`):
List of IDs.
token_ids_1 (`List[int]`, *optional*):
Optional second list of IDs for sequence pairs.
Returns:
`List[int]`: List of zeros.
Nr )r? r& )r/ r9 r: Zeosr r r $create_token_type_ids_from_sequences s z2ByT5Tokenizer.create_token_type_ids_from_sequencesc C s, | |}|dkr|S | |}|| S dS )a
Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
adding special tokens. A sequence has the following format:
- single sequence: `X `
- pair of sequences: `A B `
Args:
token_ids_0 (`List[int]`):
List of IDs to which the special tokens will be added.
token_ids_1 (`List[int]`, *optional*):
Optional second list of IDs for sequence pairs.
Returns:
`List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
N)rB )r/ r9 r: r r r build_inputs_with_special_tokens s
z.ByT5Tokenizer.build_inputs_with_special_tokens)textr c C s dd | dD }|S )zPTake as input a string and return a list of strings (tokens) for words/sub-wordsc S s g | ] }t |qS r )chrr r r r r s z+ByT5Tokenizer._tokenize..utf-8)encode)r/ rE tokensr r r _tokenize s zByT5Tokenizer._tokenizec C s$ t |dkrd}nt|| j }|S )z0Converts a token (str) in an id using the vocab.r N)r& ordr+ )r/ tokenZtoken_idr r r _convert_token_to_id s z"ByT5Tokenizer._convert_token_to_idc C s t || j }|S )z=Converts an index (integer) in a token (str) using the vocab.)rF r+ )r/ indexrL r r r _convert_id_to_token s z"ByT5Tokenizer._convert_id_to_tokenc C sh d}|D ]L}|| j kr(| j | d}n$|| jkr>|d}ntt|g}||7 }q|jddd}|S )z:Converts a sequence of tokens (string) in a single string.r rG ignore)errors)Zadded_tokens_decoderrH r7 bytesrK decode)r/ rI bstringrL Z
tok_stringstringr r r convert_tokens_to_string s
z&ByT5Tokenizer.convert_tokens_to_string)save_directoryfilename_prefixr c C s dS )Nr r )r/ rW rX r r r save_vocabulary s z ByT5Tokenizer.save_vocabulary)r
r r r
N)NF)N)N)N)__name__
__module__ __qualname____doc__Zmodel_input_namesr. propertyr4 r8 r intr r r< rB rC rD r rJ rM rO rV r rY
__classcell__r r r1 r r sN '
r )r] r@ typingr r r Ztokenization_utilsr r utilsr Z
get_loggerrZ loggerr r r r r s