-
Notifications
You must be signed in to change notification settings - Fork 3
/
tokenization_utils_base.py
2182 lines (1959 loc) · 97.4 KB
/
tokenization_utils_base.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# coding=utf-8
# Copyright 2020 The HuggingFace Inc. team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Base classes common to both the slow and the fast temporal tokenization classes: TempoPreTrainedTokenizerBase (host all the user
fronting encoding methods) and TempoSpecialTokensMixin (host the special tokens logic)
"""
import copy
import json
import os
import warnings
from typing import TYPE_CHECKING, Any, Dict, List, Optional, Tuple, Union
import numpy as np
import requests
from transformers.file_utils import (
PushToHubMixin,
cached_path,
hf_bucket_url,
is_remote_url,
is_tf_available,
is_torch_available,
)
from transformers.tokenization_utils_base import (
ADDED_TOKENS_FILE,
FULL_TOKENIZER_FILE,
LARGE_INTEGER,
SPECIAL_TOKENS_MAP_FILE,
TOKENIZER_CONFIG_FILE,
VERY_LARGE_INTEGER,
AddedToken,
BatchEncoding,
EncodedInput,
EncodedInputPair,
PaddingStrategy,
PreTokenizedInput,
PreTokenizedInputPair,
SpecialTokensMixin,
TensorType,
TextInput,
TextInputPair,
TruncationStrategy,
_is_tensorflow,
_is_torch,
to_py_obj,
)
from transformers.utils import logging
if TYPE_CHECKING:
if is_torch_available():
import torch
if is_tf_available():
import tensorflow as tf
logger = logging.get_logger(__name__)
class TempoSpecialTokensMixin(SpecialTokensMixin):
"""
A mixin derived by :class:`~transformers.PreTrainedTokenizer` and :class:`~transformers.PreTrainedTokenizerFast` to
handle specific behaviors related to special tokens. In particular, this class hold the attributes which can be
used to directly access these special tokens in a model-independent manner and allow to set and update the special
tokens.
Args:
bos_token (:obj:`str` or :obj:`tokenizers.AddedToken`, `optional`):
A special token representing the beginning of a sentence.
eos_token (:obj:`str` or :obj:`tokenizers.AddedToken`, `optional`):
A special token representing the end of a sentence.
unk_token (:obj:`str` or :obj:`tokenizers.AddedToken`, `optional`):
A special token representing an out-of-vocabulary token.
sep_token (:obj:`str` or :obj:`tokenizers.AddedToken`, `optional`):
A special token separating two different sentences in the same input (used by BERT for instance).
pad_token (:obj:`str` or :obj:`tokenizers.AddedToken`, `optional`):
A special token used to make arrays of tokens the same size for batching purpose. Will then be ignored by
attention mechanisms or loss computation.
cls_token (:obj:`str` or :obj:`tokenizers.AddedToken`, `optional`):
A special token representing the class of the input (used by BERT for instance).
mask_token (:obj:`str` or :obj:`tokenizers.AddedToken`, `optional`):
A special token representing a masked token (used by masked-language modeling pretraining objectives, like
BERT).
additional_special_tokens (tuple or list of :obj:`str` or :obj:`tokenizers.AddedToken`, `optional`):
A tuple or a list of additional special tokens.
"""
SPECIAL_TIMES_COUNT = 2 # The special "times" are: pad and mask
def __init__(self, verbose=True, **kwargs):
super().__init__(verbose, **kwargs)
@property
def pad_time_id(self) -> int:
"""
:obj:`int`: Id of the padding "time" in the time vocabulary.
"""
return self._pad_time_id
@property
def mask_time_id(self) -> int:
"""
:obj:`int`: Id of the mask "time" in the time vocabulary.
"""
return self._mask_time_id
class TempoPreTrainedTokenizerBase(TempoSpecialTokensMixin, PushToHubMixin):
"""
Base class for :class:`~transformers.PreTrainedTokenizer` and :class:`~transformers.PreTrainedTokenizerFast`.
Handles shared (mostly boiler plate) methods for those two classes.
"""
vocab_files_names: Dict[str, str] = {}
pretrained_vocab_files_map: Dict[str, Dict[str, str]] = {}
pretrained_init_configuration: Dict[str, Dict[str, Any]] = {}
max_model_input_sizes: Dict[str, Optional[int]] = {}
# first name has to correspond to main model input name
# to make sure `tokenizer.pad(...)` works correctly
model_input_names: List[str] = [
"input_ids",
"token_type_ids",
"attention_mask",
"time_ids",
]
padding_side: str = "right"
slow_tokenizer_class = None
def __init__(self, **kwargs):
# inputs and kwargs for saving and re-loading (see ``from_pretrained`` and ``save_pretrained``)
self.init_inputs = ()
self.init_kwargs = copy.deepcopy(kwargs)
self.name_or_path = kwargs.pop("name_or_path", "")
# For backward compatibility we fallback to set model_max_length from max_len if provided
model_max_length = kwargs.pop("model_max_length", kwargs.pop("max_len", None))
self.model_max_length = (
model_max_length if model_max_length is not None else VERY_LARGE_INTEGER
)
# Padding side is right by default and overridden in subclasses. If specified in the kwargs, it is changed.
self.padding_side = kwargs.pop("padding_side", self.padding_side)
assert self.padding_side in [
"right",
"left",
], f"Padding side should be selected between 'right' and 'left', current value: {self.padding_side}"
self.model_input_names = kwargs.pop("model_input_names", self.model_input_names)
self.deprecation_warnings = (
{}
) # Use to store when we have already noticed a deprecation warning (avoid overlogging).
times = kwargs.pop("times")
self.time_embedding_type = kwargs.pop("time_embedding_type")
self.init_times(times)
super().__init__(**kwargs)
def init_times(self, times, time_embedding_type=None):
# Reserve the first IDs to the "special" times (padding and masking)
# NOTE: the below IDs and `SPECIAL_TIMES_COUNT` are hardcoded
self.time_to_id = {
time: i + self.SPECIAL_TIMES_COUNT for i, time in enumerate(times)
}
self._pad_time_id = 0
self._mask_time_id = 1
if time_embedding_type:
self.time_embedding_type = time_embedding_type
if self.time_embedding_type.startswith("prepend_token"):
special_tokens = [f"<{time}>" for time in times]
self.add_tokens(special_tokens, special_tokens=True)
@classmethod
def from_non_temporal(
cls, tokenizer, config=None, times=None, time_embedding_type=None
):
tokenizer.__class__ = cls
if times is None:
times = config.times
if time_embedding_type is None:
time_embedding_type = config.time_embedding_type
tokenizer.model_input_names = cls.model_input_names
tokenizer.init_times(times, time_embedding_type)
if times:
# TempoBERT always uses a max length of 128 (hardcoded)
tokenizer.model_max_length = 128
return tokenizer
@property
def max_len_single_sentence(self) -> int:
"""
:obj:`int`: The maximum length of a sentence that can be fed to the model.
"""
return self.model_max_length - self.num_special_tokens_to_add(pair=False)
@property
def max_len_sentences_pair(self) -> int:
"""
:obj:`int`: The maximum combined length of a pair of sentences that can be fed to the model.
"""
return self.model_max_length - self.num_special_tokens_to_add(pair=True)
@max_len_single_sentence.setter
def max_len_single_sentence(self, value) -> int:
# For backward compatibility, allow to try to setup 'max_len_single_sentence'.
if (
value == self.model_max_length - self.num_special_tokens_to_add(pair=False)
and self.verbose
):
if not self.deprecation_warnings.get("max_len_single_sentence", False):
logger.warning(
"Setting 'max_len_single_sentence' is now deprecated. "
"This value is automatically set up."
)
self.deprecation_warnings["max_len_single_sentence"] = True
else:
raise ValueError(
"Setting 'max_len_single_sentence' is now deprecated. "
"This value is automatically set up."
)
@max_len_sentences_pair.setter
def max_len_sentences_pair(self, value) -> int:
# For backward compatibility, allow to try to setup 'max_len_sentences_pair'.
if (
value == self.model_max_length - self.num_special_tokens_to_add(pair=True)
and self.verbose
):
if not self.deprecation_warnings.get("max_len_sentences_pair", False):
logger.warning(
"Setting 'max_len_sentences_pair' is now deprecated. "
"This value is automatically set up."
)
self.deprecation_warnings["max_len_sentences_pair"] = True
else:
raise ValueError(
"Setting 'max_len_sentences_pair' is now deprecated. "
"This value is automatically set up."
)
def __repr__(self) -> str:
return (
f"{'PreTrainedTokenizerFast' if self.is_fast else 'PreTrainedTokenizer'}(name_or_path='{self.name_or_path}', "
f"vocab_size={self.vocab_size}, model_max_len={self.model_max_length}, is_fast={self.is_fast}, "
f"padding_side='{self.padding_side}', special_tokens={self.special_tokens_map_extended})"
)
def get_vocab(self) -> Dict[str, int]:
"""
Returns the vocabulary as a dictionary of token to index.
:obj:`tokenizer.get_vocab()[token]` is equivalent to :obj:`tokenizer.convert_tokens_to_ids(token)` when
:obj:`token` is in the vocab.
Returns:
:obj:`Dict[str, int]`: The vocabulary.
"""
raise NotImplementedError()
@classmethod
def from_pretrained(
cls,
pretrained_model_name_or_path: Union[str, os.PathLike],
*init_inputs,
**kwargs,
):
r"""
Instantiate a :class:`~transformers.tokenization_utils_base.PreTrainedTokenizerBase` (or a derived class) from
a predefined tokenizer.
Args:
pretrained_model_name_or_path (:obj:`str` or :obj:`os.PathLike`):
Can be either:
- A string, the `model id` of a predefined tokenizer hosted inside a model repo on huggingface.co.
Valid model ids can be located at the root-level, like ``bert-base-uncased``, or namespaced under a
user or organization name, like ``dbmdz/bert-base-german-cased``.
- A path to a `directory` containing vocabulary files required by the tokenizer, for instance saved
using the :meth:`~transformers.tokenization_utils_base.PreTrainedTokenizerBase.save_pretrained`
method, e.g., ``./my_model_directory/``.
- (**Deprecated**, not applicable to all derived classes) A path or url to a single saved vocabulary
file (if and only if the tokenizer only requires a single vocabulary file like Bert or XLNet), e.g.,
``./my_model_directory/vocab.txt``.
cache_dir (:obj:`str` or :obj:`os.PathLike`, `optional`):
Path to a directory in which a downloaded predefined tokenizer vocabulary files should be cached if the
standard cache should not be used.
force_download (:obj:`bool`, `optional`, defaults to :obj:`False`):
Whether or not to force the (re-)download the vocabulary files and override the cached versions if they
exist.
resume_download (:obj:`bool`, `optional`, defaults to :obj:`False`):
Whether or not to delete incompletely received files. Attempt to resume the download if such a file
exists.
proxies (:obj:`Dict[str, str], `optional`):
A dictionary of proxy servers to use by protocol or endpoint, e.g., :obj:`{'http': 'foo.bar:3128',
'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
revision(:obj:`str`, `optional`, defaults to :obj:`"main"`):
The specific model version to use. It can be a branch name, a tag name, or a commit id, since we use a
git-based system for storing models and other artifacts on huggingface.co, so ``revision`` can be any
identifier allowed by git.
subfolder (:obj:`str`, `optional`):
In case the relevant files are located inside a subfolder of the model repo on huggingface.co (e.g. for
facebook/rag-token-base), specify it here.
inputs (additional positional arguments, `optional`):
Will be passed along to the Tokenizer ``__init__`` method.
kwargs (additional keyword arguments, `optional`):
Will be passed to the Tokenizer ``__init__`` method. Can be used to set special tokens like
``bos_token``, ``eos_token``, ``unk_token``, ``sep_token``, ``pad_token``, ``cls_token``,
``mask_token``, ``additional_special_tokens``. See parameters in the ``__init__`` for more details.
Examples::
# We can't instantiate directly the base class `PreTrainedTokenizerBase` so let's show our examples on a derived class: BertTokenizer
# Download vocabulary from huggingface.co and cache.
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
# Download vocabulary from huggingface.co (user-uploaded) and cache.
tokenizer = BertTokenizer.from_pretrained('dbmdz/bert-base-german-cased')
# If vocabulary files are in a directory (e.g. tokenizer was saved using `save_pretrained('./test/saved_model/')`)
tokenizer = BertTokenizer.from_pretrained('./test/saved_model/')
# If the tokenizer uses a single vocabulary file, you can point directly to this file
tokenizer = BertTokenizer.from_pretrained('./test/saved_model/my_vocab.txt')
# You can link tokens to special vocabulary when instantiating
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased', unk_token='<unk>')
# You should be sure '<unk>' is in the vocabulary when doing that.
# Otherwise use tokenizer.add_special_tokens({'unk_token': '<unk>'}) instead)
assert tokenizer.unk_token == '<unk>'
"""
cache_dir = kwargs.pop("cache_dir", None)
force_download = kwargs.pop("force_download", False)
resume_download = kwargs.pop("resume_download", False)
proxies = kwargs.pop("proxies", None)
local_files_only = kwargs.pop("local_files_only", False)
revision = kwargs.pop("revision", None)
subfolder = kwargs.pop("subfolder", None)
s3_models = list(cls.max_model_input_sizes.keys())
pretrained_model_name_or_path = str(pretrained_model_name_or_path)
vocab_files = {}
init_configuration = {}
if pretrained_model_name_or_path in s3_models:
# Get the vocabulary from AWS S3 bucket
for file_id, map_list in cls.pretrained_vocab_files_map.items():
vocab_files[file_id] = map_list[pretrained_model_name_or_path]
if (
cls.pretrained_init_configuration
and pretrained_model_name_or_path in cls.pretrained_init_configuration
):
init_configuration = cls.pretrained_init_configuration[
pretrained_model_name_or_path
].copy()
else:
# Get the vocabulary from local files
logger.info(
"Model name '{}' not found in model shortcut name list ({}). "
"Assuming '{}' is a path, a model identifier, or url to a directory containing tokenizer files.".format(
pretrained_model_name_or_path,
", ".join(s3_models),
pretrained_model_name_or_path,
)
)
if os.path.isfile(pretrained_model_name_or_path) or is_remote_url(
pretrained_model_name_or_path
):
if len(cls.vocab_files_names) > 1:
raise ValueError(
f"Calling {cls.__name__}.from_pretrained() with the path to a single file or url is not "
"supported. Use a model identifier or the path to a directory instead."
)
warnings.warn(
f"Calling {cls.__name__}.from_pretrained() with the path to a single file or url is deprecated and "
"won't be possible anymore in v5. Use a model identifier or the path to a directory instead.",
FutureWarning,
)
file_id = list(cls.vocab_files_names.keys())[0]
vocab_files[file_id] = pretrained_model_name_or_path
else:
# At this point pretrained_model_name_or_path is either a directory or a model identifier name
additional_files_names = {
"added_tokens_file": ADDED_TOKENS_FILE,
"special_tokens_map_file": SPECIAL_TOKENS_MAP_FILE,
"tokenizer_config_file": TOKENIZER_CONFIG_FILE,
"tokenizer_file": FULL_TOKENIZER_FILE,
}
# Look for the tokenizer files
for file_id, file_name in {
**cls.vocab_files_names,
**additional_files_names,
}.items():
if os.path.isdir(pretrained_model_name_or_path):
if subfolder is not None:
full_file_name = os.path.join(
pretrained_model_name_or_path, subfolder, file_name
)
else:
full_file_name = os.path.join(
pretrained_model_name_or_path, file_name
)
if not os.path.exists(full_file_name):
logger.info(
f"Didn't find file {full_file_name}. We won't load it."
)
full_file_name = None
else:
full_file_name = hf_bucket_url(
pretrained_model_name_or_path,
filename=file_name,
subfolder=subfolder,
revision=revision,
mirror=None,
)
vocab_files[file_id] = full_file_name
# Get files from url, cache, or disk depending on the case
resolved_vocab_files = {}
unresolved_files = []
for file_id, file_path in vocab_files.items():
if file_path is None:
resolved_vocab_files[file_id] = None
else:
try:
try:
resolved_vocab_files[file_id] = cached_path(
file_path,
cache_dir=cache_dir,
force_download=force_download,
proxies=proxies,
resume_download=resume_download,
local_files_only=local_files_only,
)
except FileNotFoundError as error:
if local_files_only:
unresolved_files.append(file_id)
else:
raise error
except requests.exceptions.HTTPError as err:
if "404 Client Error" in str(err):
logger.debug(err)
resolved_vocab_files[file_id] = None
else:
raise err
if len(unresolved_files) > 0:
logger.info(
f"Can't load following files from cache: {unresolved_files} and cannot check if these "
"files are necessary for the tokenizer to operate."
)
if all(
full_file_name is None for full_file_name in resolved_vocab_files.values()
):
msg = (
f"Can't load tokenizer for '{pretrained_model_name_or_path}'. Make sure that:\n\n"
f"- '{pretrained_model_name_or_path}' is a correct model identifier listed on 'https://huggingface.co/models'\n\n"
f"- or '{pretrained_model_name_or_path}' is the correct path to a directory containing relevant tokenizer files\n\n"
)
raise EnvironmentError(msg)
for file_id, file_path in vocab_files.items():
if file_id not in resolved_vocab_files:
continue
if file_path == resolved_vocab_files[file_id]:
logger.info(f"loading file {file_path}")
else:
logger.info(
f"loading file {file_path} from cache at {resolved_vocab_files[file_id]}"
)
return cls._from_pretrained(
resolved_vocab_files,
pretrained_model_name_or_path,
init_configuration,
*init_inputs,
**kwargs,
)
@classmethod
def _from_pretrained(
cls,
resolved_vocab_files,
pretrained_model_name_or_path,
init_configuration,
*init_inputs,
**kwargs,
):
# We instantiate fast tokenizers based on a slow tokenizer if we don't have access to the tokenizer.json
# file or if `from_slow` is set to True.
from_slow = kwargs.get("from_slow", False)
has_tokenizer_file = (
resolved_vocab_files.get("tokenizer_file", None) is not None
)
if (
from_slow or not has_tokenizer_file
) and cls.slow_tokenizer_class is not None:
slow_tokenizer = (cls.slow_tokenizer_class)._from_pretrained(
copy.deepcopy(resolved_vocab_files),
pretrained_model_name_or_path,
copy.deepcopy(init_configuration),
*init_inputs,
**(copy.deepcopy(kwargs)),
)
else:
slow_tokenizer = None
# Prepare tokenizer initialization kwargs
# Did we saved some inputs and kwargs to reload ?
tokenizer_config_file = resolved_vocab_files.pop("tokenizer_config_file", None)
if tokenizer_config_file is not None:
with open(
tokenizer_config_file, encoding="utf-8"
) as tokenizer_config_handle:
init_kwargs = json.load(tokenizer_config_handle)
# First attempt. We get tokenizer_class from tokenizer_config to check mismatch between tokenizers.
config_tokenizer_class = init_kwargs.get("tokenizer_class")
init_kwargs.pop("tokenizer_class", None)
saved_init_inputs = init_kwargs.pop("init_inputs", ())
if not init_inputs:
init_inputs = saved_init_inputs
else:
config_tokenizer_class = None
init_kwargs = init_configuration
if config_tokenizer_class is None:
from transformers.models.auto.configuration_auto import AutoConfig
# Second attempt. If we have not yet found tokenizer_class, let's try to use the config.
try:
config = AutoConfig.from_pretrained(pretrained_model_name_or_path)
config_tokenizer_class = config.tokenizer_class
except (OSError, ValueError, KeyError):
# skip if an error occured.
config = None
if config_tokenizer_class is None:
# Third attempt. If we have not yet found the original type of the tokenizer,
# we are loading we see if we can infer it from the type of the configuration file
from transformers.models.auto.configuration_auto import CONFIG_MAPPING
from transformers.models.auto.tokenization_auto import TOKENIZER_MAPPING
if hasattr(config, "model_type"):
config_class = CONFIG_MAPPING.get(config.model_type)
else:
# Fallback: use pattern matching on the string.
config_class = None
for pattern, config_class_tmp in CONFIG_MAPPING.items():
if pattern in str(pretrained_model_name_or_path):
config_class = config_class_tmp
break
if config_class in TOKENIZER_MAPPING.keys():
(
config_tokenizer_class,
config_tokenizer_class_fast,
) = TOKENIZER_MAPPING[config_class]
if config_tokenizer_class is not None:
config_tokenizer_class = config_tokenizer_class.__name__
else:
config_tokenizer_class = config_tokenizer_class_fast.__name__
if config_tokenizer_class is not None:
if cls.__name__.replace("Fast", "") != config_tokenizer_class.replace(
"Fast", ""
):
logger.warning(
"The tokenizer class you load from this checkpoint is not the same type as the class this function is called from. "
"It may result in unexpected tokenization. \n"
f"The tokenizer class you load from this checkpoint is '{config_tokenizer_class}'. \n"
f"The class this function is called from is '{cls.__name__}'."
)
# Update with newly provided kwargs
init_kwargs.update(kwargs)
# Convert AddedTokens serialized as dict to class instances
def convert_added_tokens(obj: Union[AddedToken, Any]):
if (
isinstance(obj, dict)
and "__type" in obj
and obj["__type"] == "AddedToken"
):
obj.pop("__type")
return AddedToken(**obj)
elif isinstance(obj, (list, tuple)):
return list(convert_added_tokens(o) for o in obj)
elif isinstance(obj, dict):
return {k: convert_added_tokens(v) for k, v in obj.items()}
return obj
init_kwargs = convert_added_tokens(init_kwargs)
# Set max length if needed
if pretrained_model_name_or_path in cls.max_model_input_sizes:
# if we're using a pretrained model, ensure the tokenizer
# wont index sequences longer than the number of positional embeddings
model_max_length = cls.max_model_input_sizes[pretrained_model_name_or_path]
if model_max_length is not None and isinstance(
model_max_length, (int, float)
):
init_kwargs["model_max_length"] = min(
init_kwargs.get("model_max_length", int(1e30)), model_max_length
)
# Merge resolved_vocab_files arguments in init_kwargs.
added_tokens_file = resolved_vocab_files.pop("added_tokens_file", None)
for args_name, file_path in resolved_vocab_files.items():
if args_name not in init_kwargs:
init_kwargs[args_name] = file_path
if slow_tokenizer is not None:
init_kwargs["__slow_tokenizer"] = slow_tokenizer
init_kwargs["name_or_path"] = pretrained_model_name_or_path
# Instantiate tokenizer.
try:
tokenizer = cls(*init_inputs, **init_kwargs)
except OSError:
raise OSError(
"Unable to load vocabulary from file. "
"Please check that the provided vocabulary is accessible and not corrupted."
)
# Save inputs and kwargs for saving and re-loading with ``save_pretrained``
# Removed: Now done at the base class level
# tokenizer.init_inputs = init_inputs
# tokenizer.init_kwargs = init_kwargs
# If there is a complementary special token map, load it
special_tokens_map_file = resolved_vocab_files.pop(
"special_tokens_map_file", None
)
if special_tokens_map_file is not None:
with open(
special_tokens_map_file, encoding="utf-8"
) as special_tokens_map_handle:
special_tokens_map = json.load(special_tokens_map_handle)
for key, value in special_tokens_map.items():
if isinstance(value, dict):
value = AddedToken(**value)
elif isinstance(value, list):
value = [
AddedToken(**token) if isinstance(token, dict) else token
for token in value
]
setattr(tokenizer, key, value)
# Add supplementary tokens.
special_tokens = tokenizer.all_special_tokens
if added_tokens_file is not None:
with open(added_tokens_file, encoding="utf-8") as added_tokens_handle:
added_tok_encoder = json.load(added_tokens_handle)
# Sort added tokens by index
added_tok_encoder_sorted = list(
sorted(added_tok_encoder.items(), key=lambda x: x[1])
)
for token, index in added_tok_encoder_sorted:
if (
has_tokenizer_file
and index != len(tokenizer)
and tokenizer.convert_tokens_to_ids(token) != index
):
# Tokenizer fast: added token needs to either be in the vocabulary with the proper index or the
# index is the current length of the tokenizer (not in vocabulary)
raise ValueError(
f"Wrong index found for {token}: should be {tokenizer.convert_tokens_to_ids(token)} but found "
f"{index}."
)
elif not has_tokenizer_file and index != len(tokenizer):
# Tokenizer slow: added token cannot already be in the vocabulary so its index needs to be the
# current length of the tokenizer.
raise ValueError(
f"Non-consecutive added token '{token}' found. "
f"Should have index {len(tokenizer)} but has index {index} in saved vocabulary."
)
# Safe to call on a tokenizer fast even if token already there.
tokenizer.add_tokens(
token, special_tokens=bool(token in special_tokens)
)
# Check all our special tokens are registered as "no split" token (we don't cut them) and are in the vocab
added_tokens = tokenizer.sanitize_special_tokens()
if added_tokens:
logger.warning(
"Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained."
)
return tokenizer
def save_pretrained(
self,
save_directory: Union[str, os.PathLike],
legacy_format: Optional[bool] = None,
filename_prefix: Optional[str] = None,
push_to_hub: bool = False,
**kwargs,
) -> Tuple[str]:
"""
Save the full tokenizer state.
This method make sure the full tokenizer can then be re-loaded using the
:meth:`~transformers.tokenization_utils_base.PreTrainedTokenizer.from_pretrained` class method..
.. Warning::
This won't save modifications you may have applied to the tokenizer after the instantiation (for instance,
modifying :obj:`tokenizer.do_lower_case` after creation).
Args:
save_directory (:obj:`str` or :obj:`os.PathLike`): The path to a directory where the tokenizer will be saved.
legacy_format (:obj:`bool`, `optional`):
Only applicable for a fast tokenizer. If unset (default), will save the tokenizer in the unified JSON
format as well as in legacy format if it exists, i.e. with tokenizer specific vocabulary and a separate
added_tokens files.
If :obj:`False`, will only save the tokenizer in the unified JSON format. This format is incompatible
with "slow" tokenizers (not powered by the `tokenizers` library), so the tokenizer will not be able to
be loaded in the corresponding "slow" tokenizer.
If :obj:`True`, will save the tokenizer in legacy format. If the "slow" tokenizer doesn't exits, a
value error is raised.
filename_prefix: (:obj:`str`, `optional`):
A prefix to add to the names of the files saved by the tokenizer.
Returns:
A tuple of :obj:`str`: The files saved.
"""
if os.path.isfile(save_directory):
logger.error(
f"Provided path ({save_directory}) should be a directory, not a file"
)
return
os.makedirs(save_directory, exist_ok=True)
special_tokens_map_file = os.path.join(
save_directory,
(filename_prefix + "-" if filename_prefix else "")
+ SPECIAL_TOKENS_MAP_FILE,
)
tokenizer_config_file = os.path.join(
save_directory,
(filename_prefix + "-" if filename_prefix else "") + TOKENIZER_CONFIG_FILE,
)
tokenizer_config = copy.deepcopy(self.init_kwargs)
if len(self.init_inputs) > 0:
tokenizer_config["init_inputs"] = copy.deepcopy(self.init_inputs)
for file_id in self.vocab_files_names.keys():
tokenizer_config.pop(file_id, None)
# Sanitize AddedTokens
def convert_added_tokens(obj: Union[AddedToken, Any], add_type_field=True):
if isinstance(obj, AddedToken):
out = obj.__getstate__()
if add_type_field:
out["__type"] = "AddedToken"
return out
elif isinstance(obj, (list, tuple)):
return list(
convert_added_tokens(o, add_type_field=add_type_field) for o in obj
)
elif isinstance(obj, dict):
return {
k: convert_added_tokens(v, add_type_field=add_type_field)
for k, v in obj.items()
}
return obj
# add_type_field=True to allow dicts in the kwargs / differentiate from AddedToken serialization
tokenizer_config = convert_added_tokens(tokenizer_config, add_type_field=True)
# Add tokenizer class to the tokenizer config to be able to reload it with from_pretrained
tokenizer_class = self.__class__.__name__
# Remove the Fast at the end unless we have a special `PreTrainedTokenizerFast`
if (
tokenizer_class.endswith("Fast")
and tokenizer_class != "PreTrainedTokenizerFast"
):
tokenizer_class = tokenizer_class[:-4]
tokenizer_config["tokenizer_class"] = tokenizer_class
with open(tokenizer_config_file, "w", encoding="utf-8") as f:
f.write(json.dumps(tokenizer_config, ensure_ascii=False))
# Sanitize AddedTokens in special_tokens_map
write_dict = convert_added_tokens(
self.special_tokens_map_extended, add_type_field=False
)
with open(special_tokens_map_file, "w", encoding="utf-8") as f:
f.write(json.dumps(write_dict, ensure_ascii=False))
file_names = (tokenizer_config_file, special_tokens_map_file)
save_files = self._save_pretrained(
save_directory=save_directory,
file_names=file_names,
legacy_format=legacy_format,
filename_prefix=filename_prefix,
)
if push_to_hub:
# Annoyingly, the return contains files that don't exist.
existing_files = [f for f in save_files if os.path.isfile(f)]
url = self._push_to_hub(save_files=existing_files, **kwargs)
logger.info(f"Tokenizer pushed to the hub in this commit: {url}")
return save_files
def _save_pretrained(
self,
save_directory: Union[str, os.PathLike],
file_names: Tuple[str],
legacy_format: Optional[bool] = None,
filename_prefix: Optional[str] = None,
) -> Tuple[str]:
"""
Save a tokenizer using the slow-tokenizer/legacy format: vocabulary + added tokens.
Fast tokenizers can also be saved in a unique JSON file containing {config + vocab + added-tokens} using the
specific :meth:`~transformers.tokenization_utils_fast.PreTrainedTokenizerFast._save_pretrained`
"""
if legacy_format is False:
raise ValueError(
"Only fast tokenizers (instances of PreTrainedTokenizerFast) can be saved in non legacy format."
)
save_directory = str(save_directory)
added_tokens_file = os.path.join(
save_directory,
(filename_prefix + "-" if filename_prefix else "") + ADDED_TOKENS_FILE,
)
added_vocab = self.get_added_vocab()
if added_vocab:
with open(added_tokens_file, "w", encoding="utf-8") as f:
out_str = json.dumps(added_vocab, ensure_ascii=False)
f.write(out_str)
vocab_files = self.save_vocabulary(
save_directory, filename_prefix=filename_prefix
)
return file_names + vocab_files + (added_tokens_file,)
def save_vocabulary(
self, save_directory: str, filename_prefix: Optional[str] = None
) -> Tuple[str]:
"""
Save only the vocabulary of the tokenizer (vocabulary + added tokens).
This method won't save the configuration and special token mappings of the tokenizer. Use
:meth:`~transformers.PreTrainedTokenizerFast._save_pretrained` to save the whole state of the tokenizer.
Args:
save_directory (:obj:`str`):
The directory in which to save the vocabulary.
filename_prefix (:obj:`str`, `optional`):
An optional prefix to add to the named of the saved files.
Returns:
:obj:`Tuple(str)`: Paths to the files saved.
"""
raise NotImplementedError
def tokenize(
self,
text: str,
pair: Optional[str] = None,
add_special_tokens: bool = False,
**kwargs,
) -> List[str]:
"""
Converts a string in a sequence of tokens, replacing unknown tokens with the :obj:`unk_token`.
Args:
text (:obj:`str`):
The sequence to be encoded.
pair (:obj:`str`, `optional`):
A second sequence to be encoded with the first.
add_special_tokens (:obj:`bool`, `optional`, defaults to :obj:`False`):
Whether or not to add the special tokens associated with the corresponding model.
kwargs (additional keyword arguments, `optional`):
Will be passed to the underlying model specific encode method. See details in
:meth:`~transformers.PreTrainedTokenizerBase.__call__`
Returns:
:obj:`List[str]`: The list of tokens.
"""
raise NotImplementedError
def encode(
self,
text: Union[TextInput, PreTokenizedInput, EncodedInput],
text_pair: Optional[Union[TextInput, PreTokenizedInput, EncodedInput]] = None,
add_special_tokens: bool = True,
padding: Union[bool, str, PaddingStrategy] = False,
truncation: Union[bool, str, TruncationStrategy] = False,
max_length: Optional[int] = None,
stride: int = 0,
return_tensors: Optional[Union[str, TensorType]] = None,
**kwargs,
) -> List[int]:
"""
Converts a string to a sequence of ids (integer), using the tokenizer and vocabulary.
Same as doing ``self.convert_tokens_to_ids(self.tokenize(text))``.
Args:
text (:obj:`str`, :obj:`List[str]` or :obj:`List[int]`):
The first sequence to be encoded. This can be a string, a list of strings (tokenized string using the
``tokenize`` method) or a list of integers (tokenized string ids using the ``convert_tokens_to_ids``
method).
text_pair (:obj:`str`, :obj:`List[str]` or :obj:`List[int]`, `optional`):
Optional second sequence to be encoded. This can be a string, a list of strings (tokenized string using
the ``tokenize`` method) or a list of integers (tokenized string ids using the
``convert_tokens_to_ids`` method).
"""
encoded_inputs = self.encode_plus(
text,
text_pair=text_pair,
add_special_tokens=add_special_tokens,
padding=padding,
truncation=truncation,
max_length=max_length,
stride=stride,
return_tensors=return_tensors,
**kwargs,
)
return encoded_inputs["input_ids"]
def num_special_tokens_to_add(self, pair: bool = False) -> int:
raise NotImplementedError
def _get_padding_truncation_strategies(
self,
padding=False,
truncation=False,
max_length=None,
pad_to_multiple_of=None,
verbose=True,
**kwargs,
):
"""
Find the correct padding/truncation strategy with backward compatibility for old arguments (truncation_strategy
and pad_to_max_length) and behaviors.
"""
old_truncation_strategy = kwargs.pop("truncation_strategy", "do_not_truncate")
old_pad_to_max_length = kwargs.pop("pad_to_max_length", False)
# Backward compatibility for previous behavior, maybe we should deprecate it:
# If you only set max_length, it activates truncation for max_length
if max_length is not None and padding is False and truncation is False:
if verbose:
if not self.deprecation_warnings.get(
"Truncation-not-explicitly-activated", False
):
logger.warning(
"Truncation was not explicitly activated but `max_length` is provided a specific value, "
"please use `truncation=True` to explicitly truncate examples to max length. "
"Defaulting to 'longest_first' truncation strategy. "
"If you encode pairs of sequences (GLUE-style) with the tokenizer you can select this strategy "
"more precisely by providing a specific strategy to `truncation`."
)
self.deprecation_warnings["Truncation-not-explicitly-activated"] = True
truncation = "longest_first"
# Get padding strategy
if padding is False and old_pad_to_max_length:
if verbose:
warnings.warn(
"The `pad_to_max_length` argument is deprecated and will be removed in a future version, "
"use `padding=True` or `padding='longest'` to pad to the longest sequence in the batch, or "
"use `padding='max_length'` to pad to a max length. In this case, you can give a specific "
"length with `max_length` (e.g. `max_length=45`) or leave max_length to None to pad to the "
"maximal input size of the model (e.g. 512 for Bert).",
FutureWarning,
)
if max_length is None:
padding_strategy = PaddingStrategy.LONGEST
else:
padding_strategy = PaddingStrategy.MAX_LENGTH
elif padding is not False: