o
Zho- @ s d dl Z d dlmZ d dlmZmZmZ d dlmZ ddl m
Z
m Z ddl m
Z
ddlmZmZ e r:d d
lmZ ndZeeZd d d
Zg dZG dd de
ZdgZdS ) N)copyfile)ListOptionalTuple)
processors )
AddedToken
BatchEncoding)PreTrainedTokenizerFast)is_sentencepiece_availablelogging )MBart50Tokenizerzsentencepiece.bpe.modelztokenizer.json)
vocab_filetokenizer_file)4Zar_ARcs_CZde_DEen_XXZes_XXet_EEfi_FIZfr_XXgu_INhi_INit_ITZja_XXkk_KZko_KRlt_LTlv_LVZmy_MMne_NPZnl_XXro_ROru_RUsi_LKtr_TRvi_VNzh_CNaf_ZAaz_AZbn_INfa_IRhe_ILhr_HRid_IDka_GEZkm_KHmk_MKml_INmn_MNmr_INpl_PLps_AFZpt_XXsv_SEsw_KEta_INte_INth_THZtl_XXuk_UAur_PKxh_ZAgl_ESsl_SIc sv e Zd ZU dZeZddgZeZg Z e
e e d< g Z
e
e e d<
d0 fd d
ZedefddZedefddZejdeddfddZ d1de
e dee
e de
e fddZ d2de
e d ed ee
e d edef
fd d
Zd!d" Zd#d$ Zd eddfd%d&Zd eddfd'd(Z d)ed ee d ee fd*d+Z d1d,ed-ee de e fd.d/Z Z S )3MBart50TokenizerFastu
Construct a "fast" MBART tokenizer for mBART-50 (backed by HuggingFace's *tokenizers* library). Based on
[BPE](https://huggingface.co/docs/tokenizers/python/latest/components.html?highlight=BPE#models).
This tokenizer inherits from [`PreTrainedTokenizerFast`] which contains most of the main methods. Users should
refer to this superclass for more information regarding those methods.
Args:
vocab_file (`str`):
Path to the vocabulary file.
src_lang (`str`, *optional*):
A string representing the source language.
tgt_lang (`str`, *optional*):
A string representing the target language.
eos_token (`str`, *optional*, defaults to `""`):
The end of sequence token.
sep_token (`str`, *optional*, defaults to `""`):
The separator token, which is used when building a sequence from multiple sequences, e.g. two sequences for
sequence classification or for a text and a question for question answering. It is also used as the last
token of a sequence built with special tokens.
cls_token (`str`, *optional*, defaults to `""`):
The classifier token which is used when doing sequence classification (classification of the whole sequence
instead of per-token classification). It is the first token of the sequence when built with special tokens.
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.
mask_token (`str`, *optional*, defaults to `""`):
The token used for masking values. This is the token used when training this model with masked language
modeling. This is the token which the model will try to predict.
Examples:
```python
>>> from transformers import MBart50TokenizerFast
>>> tokenizer = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50", src_lang="en_XX", tgt_lang="ro_RO")
>>> src_text = " UN Chief Says There Is No Military Solution in Syria"
>>> tgt_text = "Şeful ONU declară că nu există o soluţie militară în Siria"
>>> model_inputs = tokenizer(src_text, text_target=tgt_text, return_tensors="pt")
>>> # model(**model_inputs) should work
```Z input_idsZattention_mask
prefix_tokens
suffix_tokensNc s t |
tr t|
dddn|
}
dg pg d< d fddtD 7 < t j|f|||||||| |
d |_fdd tD _|d urN|nd
_ |_
jj _ j d S ) NTF)lstriprstripadditional_special_tokensc s g | ]
}| d vr|qS )rF ).0code)kwargsrG d/var/www/auris/lib/python3.10/site-packages/transformers/models/mbart50/tokenization_mbart50_fast.py
o s z1MBart50TokenizerFast.__init__..) src_langtgt_langr eos_token sep_token cls_token unk_token pad_token
mask_tokenc s i | ]}| |qS rG )convert_tokens_to_ids)rH Z lang_codeselfrG rK
s z1MBart50TokenizerFast.__init__..r )
isinstancestrr getFAIRSEQ_LANGUAGE_CODESsuper__init__r Zlang_code_to_id _src_langrN cur_lang_code_idset_src_lang_special_tokens) rW r rM rN r rO rP rQ rR rS rT rJ __class__)rJ rW rK r^ ] s8
z MBart50TokenizerFast.__init__returnc C s | j r
tj| j S dS )NF)r ospathisfilerV rG rG rK can_save_slow_tokenizer s z,MBart50TokenizerFast.can_save_slow_tokenizerc C s | j S N)r_ rV rG rG rK rM s z MBart50TokenizerFast.src_lang new_src_langc C s || _ | | j d S ri )r_ ra )rW rj rG rG rK rM s token_ids_0 token_ids_1c C s, |du r | j | | j S | j | | | j S )a
Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
adding special tokens. The special tokens depend on calling set_lang.
An MBART-50 sequence has the following format, where `X` represents the sequence:
- `input_ids` (for encoder) `[src_lang_code] X [eos]`
- `labels`: (for decoder) `[tgt_lang_code] X [eos]`
BOS is never used. Pairs of sequences are not the expected use case, but they will be handled without a
separator.
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)r= r> )rW rk rl rG rG rK build_inputs_with_special_tokens s z5MBart50TokenizerFast.build_inputs_with_special_tokensr r src_textsrM tgt_textsrN c s" || _ || _t j||fi |S ri )rM rN r] prepare_seq2seq_batch)rW rn rM ro rN rJ rb rG rK rp s z*MBart50TokenizerFast.prepare_seq2seq_batchc C | | jS ri )ra rM rV rG rG rK _switch_to_input_mode z*MBart50TokenizerFast._switch_to_input_modec C rq ri )set_tgt_lang_special_tokensrN rV rG rG rK _switch_to_target_mode rs z+MBart50TokenizerFast._switch_to_target_modec C | | || _| jg| _| jg| _| | j}| | j}tj|dg | |ddg | tt || | j| j d| j
_ dS )z]Reset the special tokens to the source lang setting. prefix=[src_lang_code] and suffix=[eos].$A$BsinglepairZspecial_tokensN rU r` r= Z eos_token_idr> Zconvert_ids_to_tokensr ZTemplateProcessinglistzipZ
_tokenizerZpost_processor)rW rM prefix_tokens_strsuffix_tokens_strrG rG rK ra
z0MBart50TokenizerFast.set_src_lang_special_tokensc C rv )zaReset the special tokens to the target language setting. prefix=[src_lang_code] and suffix=[eos].rw rx ry Nr| )rW rN r r rG rG rK rt r z0MBart50TokenizerFast.set_tgt_lang_special_tokensreturn_tensorsc K sJ |du s|du r t d|| _| |fd|d|}| |}||d< |S )zIUsed by translation pipeline, to prepare inputs for the generate functionNzATranslation requires a `src_lang` and a `tgt_lang` for this modelT)Zadd_special_tokensr Zforced_bos_token_id)
ValueErrorrM rU )rW Z
raw_inputsr rM rN extra_kwargsZinputsZ tgt_lang_idrG rG rK _build_translation_inputs s
z.MBart50TokenizerFast._build_translation_inputssave_directoryfilename_prefixc C s~ | j stdtj|std| d d S tj||r"|d ndtd }tj | j
tj |kr r^ propertyboolrh rZ rM setterr rm r rp rr ru ra rt r r r
__classcell__rG rG rb rK r< ) st
,/
( r< )re shutilr typingr r r Z
tokenizersr Ztokenization_utilsr r Ztokenization_utils_fastr
utilsr r Ztokenization_mbart50r Z
get_loggerr r r r\ r< __all__rG rG rG rK s
^