You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
1032 lines
48 KiB
1032 lines
48 KiB
# Copyright 2024 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
|
|
"""Tokenization classes for UDOP model."""
|
|
|
|
from tokenizers import Tokenizer, decoders, pre_tokenizers, processors
|
|
from tokenizers.models import Unigram
|
|
|
|
from ...tokenization_utils_base import (
|
|
BatchEncoding,
|
|
EncodedInput,
|
|
PreTokenizedInput,
|
|
TextInput,
|
|
TextInputPair,
|
|
TruncationStrategy,
|
|
)
|
|
from ...tokenization_utils_tokenizers import TokenizersBackend
|
|
from ...utils import PaddingStrategy, TensorType, add_end_docstrings, logging
|
|
|
|
|
|
VOCAB_FILES_NAMES = {"vocab_file": "spiece.model", "tokenizer_file": "tokenizer.json"}
|
|
|
|
|
|
logger = logging.get_logger(__name__)
|
|
|
|
UDOP_ENCODE_KWARGS_DOCSTRING = r"""
|
|
add_special_tokens (`bool`, *optional*, defaults to `True`):
|
|
Whether or not to encode the sequences with the special tokens relative to their model.
|
|
padding (`bool`, `str` or [`~file_utils.PaddingStrategy`], *optional*, defaults to `False`):
|
|
Activates and controls padding. Accepts the following values:
|
|
|
|
- `True` or `'longest'`: Pad to the longest sequence in the batch (or no padding if only a single
|
|
sequence if provided).
|
|
- `'max_length'`: Pad to a maximum length specified with the argument `max_length` or to the maximum
|
|
acceptable input length for the model if that argument is not provided.
|
|
- `False` or `'do_not_pad'` (default): No padding (i.e., can output a batch with sequences of different
|
|
lengths).
|
|
truncation (`bool`, `str` or [`~tokenization_utils_base.TruncationStrategy`], *optional*, defaults to `False`):
|
|
Activates and controls truncation. Accepts the following values:
|
|
|
|
- `True` or `'longest_first'`: Truncate to a maximum length specified with the argument `max_length` or
|
|
to the maximum acceptable input length for the model if that argument is not provided. This will
|
|
truncate token by token, removing a token from the longest sequence in the pair if a pair of
|
|
sequences (or a batch of pairs) is provided.
|
|
- `'only_first'`: Truncate to a maximum length specified with the argument `max_length` or to the
|
|
maximum acceptable input length for the model if that argument is not provided. This will only
|
|
truncate the first sequence of a pair if a pair of sequences (or a batch of pairs) is provided.
|
|
- `'only_second'`: Truncate to a maximum length specified with the argument `max_length` or to the
|
|
maximum acceptable input length for the model if that argument is not provided. This will only
|
|
truncate the second sequence of a pair if a pair of sequences (or a batch of pairs) is provided.
|
|
- `False` or `'do_not_truncate'` (default): No truncation (i.e., can output batch with sequence lengths
|
|
greater than the model maximum admissible input size).
|
|
max_length (`int`, *optional*):
|
|
Controls the maximum length to use by one of the truncation/padding parameters.
|
|
|
|
If left unset or set to `None`, this will use the predefined model maximum length if a maximum length
|
|
is required by one of the truncation/padding parameters. If the model has no specific maximum input
|
|
length (like XLNet) truncation/padding to a maximum length will be deactivated.
|
|
stride (`int`, *optional*, defaults to 0):
|
|
If set to a number along with `max_length`, the overflowing tokens returned when
|
|
`return_overflowing_tokens=True` will contain some tokens from the end of the truncated sequence
|
|
returned to provide some overlap between truncated and overflowing sequences. The value of this
|
|
argument defines the number of overlapping tokens.
|
|
pad_to_multiple_of (`int`, *optional*):
|
|
If set will pad the sequence to a multiple of the provided value. This is especially useful to enable
|
|
the use of Tensor Cores on NVIDIA hardware with compute capability `>= 7.5` (Volta).
|
|
return_tensors (`str` or [`~file_utils.TensorType`], *optional*):
|
|
If set, will return tensors instead of list of python integers. Acceptable values are:
|
|
|
|
- `'pt'`: Return PyTorch `torch.Tensor` objects.
|
|
- `'np'`: Return Numpy `np.ndarray` objects.
|
|
return_token_type_ids (`bool`, *optional*):
|
|
Whether to return token type IDs. If left to the default, will return the token type IDs according to
|
|
the specific tokenizer's default, defined by the `return_outputs` attribute.
|
|
|
|
[What are token type IDs?](../glossary#token-type-ids)
|
|
return_attention_mask (`bool`, *optional*):
|
|
Whether to return the attention mask. If left to the default, will return the attention mask according
|
|
to the specific tokenizer's default, defined by the `return_outputs` attribute.
|
|
|
|
[What are attention masks?](../glossary#attention-mask)
|
|
return_overflowing_tokens (`bool`, *optional*, defaults to `False`):
|
|
Whether or not to return overflowing token sequences. If a pair of sequences of input ids (or a batch
|
|
of pairs) is provided with `truncation_strategy = longest_first` or `True`, an error is raised instead
|
|
of returning overflowing tokens.
|
|
return_special_tokens_mask (`bool`, *optional*, defaults to `False`):
|
|
Whether or not to return special tokens mask information.
|
|
return_offsets_mapping (`bool`, *optional*, defaults to `False`):
|
|
Whether or not to return `(char_start, char_end)` for each token.
|
|
|
|
This is only available on fast tokenizers inheriting from [`PreTrainedTokenizerFast`], if using
|
|
Python's tokenizer, this method will raise `NotImplementedError`.
|
|
return_length (`bool`, *optional*, defaults to `False`):
|
|
Whether or not to return the lengths of the encoded inputs.
|
|
verbose (`bool`, *optional*, defaults to `True`):
|
|
Whether or not to print more information and warnings.
|
|
**kwargs: passed to the `self.tokenize()` method
|
|
|
|
Return:
|
|
[`BatchEncoding`]: A [`BatchEncoding`] with the following fields:
|
|
|
|
- **input_ids** -- List of token ids to be fed to a model.
|
|
|
|
[What are input IDs?](../glossary#input-ids)
|
|
|
|
- **bbox** -- List of bounding boxes to be fed to a model.
|
|
|
|
- **token_type_ids** -- List of token type ids to be fed to a model (when `return_token_type_ids=True` or
|
|
if *"token_type_ids"* is in `self.model_input_names`).
|
|
|
|
[What are token type IDs?](../glossary#token-type-ids)
|
|
|
|
- **attention_mask** -- List of indices specifying which tokens should be attended to by the model (when
|
|
`return_attention_mask=True` or if *"attention_mask"* is in `self.model_input_names`).
|
|
|
|
[What are attention masks?](../glossary#attention-mask)
|
|
|
|
- **labels** -- List of labels to be fed to a model. (when `word_labels` is specified).
|
|
- **overflowing_tokens** -- List of overflowing tokens sequences (when a `max_length` is specified and
|
|
`return_overflowing_tokens=True`).
|
|
- **num_truncated_tokens** -- Number of tokens truncated (when a `max_length` is specified and
|
|
`return_overflowing_tokens=True`).
|
|
- **special_tokens_mask** -- List of 0s and 1s, with 1 specifying added special tokens and 0 specifying
|
|
regular sequence tokens (when `add_special_tokens=True` and `return_special_tokens_mask=True`).
|
|
- **length** -- The length of the inputs (when `return_length=True`).
|
|
"""
|
|
|
|
|
|
class UdopTokenizer(TokenizersBackend):
|
|
"""
|
|
Construct a "fast" UDOP tokenizer (backed by HuggingFace's *tokenizers* library). Adapted from
|
|
[`LayoutXLMTokenizer`] and [`T5Tokenizer`]. Based on
|
|
[BPE](https://huggingface.co/docs/tokenizers/python/latest/components.html?highlight=BPE#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:
|
|
eos_token (`str`, *optional*, defaults to `"</s>"`):
|
|
The end of sequence token.
|
|
|
|
<Tip>
|
|
|
|
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`.
|
|
|
|
</Tip>
|
|
|
|
sep_token (`str`, *optional*, defaults to `"</s>"`):
|
|
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.
|
|
unk_token (`str`, *optional*, defaults to `"<unk>"`):
|
|
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 `"<pad>"`):
|
|
The token used for padding, for example when batching sequences of different lengths.
|
|
sep_token_box (`list[int]`, *optional*, defaults to `[1000, 1000, 1000, 1000]`):
|
|
The bounding box to use for the special [SEP] token.
|
|
pad_token_box (`list[int]`, *optional*, defaults to `[0, 0, 0, 0]`):
|
|
The bounding box to use for the special [PAD] token.
|
|
pad_token_label (`int`, *optional*, defaults to -100):
|
|
The label to use for padding tokens. Defaults to -100, which is the `ignore_index` of PyTorch's
|
|
CrossEntropyLoss.
|
|
only_label_first_subword (`bool`, *optional*, defaults to `True`):
|
|
Whether or not to only label the first subword, in case word labels are provided.
|
|
extra_special_tokens (`list[str]`, *optional*, defaults to `["<s>NOTUSED", "</s>NOTUSED"]`):
|
|
Extra special tokens used by the tokenizer.
|
|
"""
|
|
|
|
vocab_files_names = VOCAB_FILES_NAMES
|
|
model_input_names = ["input_ids", "attention_mask"]
|
|
model = Unigram
|
|
|
|
def __init__(
|
|
self,
|
|
vocab: str | list[tuple[str, float]] | None = None,
|
|
eos_token="</s>",
|
|
sep_token="</s>",
|
|
unk_token="<unk>",
|
|
pad_token="<pad>",
|
|
sep_token_box=[1000, 1000, 1000, 1000],
|
|
pad_token_box=[0, 0, 0, 0],
|
|
pad_token_label=-100,
|
|
only_label_first_subword=True,
|
|
extra_special_tokens=None,
|
|
**kwargs,
|
|
):
|
|
if "additional_special_tokens" in kwargs and "extra_special_tokens" not in kwargs:
|
|
kwargs["extra_special_tokens"] = kwargs.pop("additional_special_tokens")
|
|
if extra_special_tokens is not None:
|
|
kwargs["extra_special_tokens"] = extra_special_tokens
|
|
|
|
if vocab is None:
|
|
vocab = [(str(pad_token), 0.0), (str(eos_token), 0.0), (str(unk_token), 0.0), ("▁", -2.0)]
|
|
|
|
unk_id = 2
|
|
for idx, (token, _) in enumerate(vocab):
|
|
if token == str(unk_token):
|
|
unk_id = idx
|
|
break
|
|
|
|
self._tokenizer = Tokenizer(
|
|
Unigram(
|
|
vocab,
|
|
unk_id=unk_id,
|
|
byte_fallback=False,
|
|
)
|
|
)
|
|
|
|
self._tokenizer.normalizer = None
|
|
|
|
self._tokenizer.pre_tokenizer = pre_tokenizers.Sequence(
|
|
[
|
|
pre_tokenizers.WhitespaceSplit(),
|
|
pre_tokenizers.Metaspace(replacement="▁", prepend_scheme="always", split=True),
|
|
]
|
|
)
|
|
|
|
self._tokenizer.decoder = decoders.Metaspace(replacement="▁", prepend_scheme="always", split=True)
|
|
|
|
super().__init__(
|
|
eos_token=eos_token,
|
|
sep_token=sep_token,
|
|
unk_token=unk_token,
|
|
pad_token=pad_token,
|
|
sep_token_box=sep_token_box,
|
|
pad_token_box=pad_token_box,
|
|
pad_token_label=pad_token_label,
|
|
only_label_first_subword=only_label_first_subword,
|
|
**kwargs,
|
|
)
|
|
|
|
self._tokenizer.post_processor = processors.TemplateProcessing(
|
|
single=["$A", "</s>"],
|
|
pair=["$A", "</s>", "$B", "</s>"],
|
|
special_tokens=[
|
|
("</s>", self.eos_token_id),
|
|
],
|
|
)
|
|
|
|
self.sep_token_box = sep_token_box
|
|
self.pad_token_box = pad_token_box
|
|
self.pad_token_label = pad_token_label
|
|
self.only_label_first_subword = only_label_first_subword
|
|
self.init_kwargs["vocab"] = vocab
|
|
|
|
self._tokenizer.encode_special_tokens = self.split_special_tokens
|
|
|
|
@add_end_docstrings(UDOP_ENCODE_KWARGS_DOCSTRING)
|
|
def __call__(
|
|
self,
|
|
text: TextInput | PreTokenizedInput | list[TextInput] | list[PreTokenizedInput] = None,
|
|
text_pair: PreTokenizedInput | list[PreTokenizedInput] | None = None,
|
|
boxes: list[list[int]] | list[list[list[int]]] | None = None,
|
|
word_labels: list[int] | list[list[int]] | None = None,
|
|
text_target: TextInput | PreTokenizedInput | list[TextInput] | list[PreTokenizedInput] = None,
|
|
text_pair_target: TextInput | PreTokenizedInput | list[TextInput] | list[PreTokenizedInput] | None = None,
|
|
**kwargs,
|
|
) -> BatchEncoding:
|
|
if text is None and text_target is None:
|
|
raise ValueError("You need to specify either `text` or `text_target`.")
|
|
if text is not None:
|
|
# The context manager will send the inputs as normal texts and not text_target, but we shouldn't change the
|
|
# input mode in this case.
|
|
if not self._in_target_context_manager and hasattr(self, "_switch_to_input_mode"):
|
|
self._switch_to_input_mode()
|
|
encodings = self.call_boxes(text=text, text_pair=text_pair, boxes=boxes, word_labels=word_labels, **kwargs)
|
|
if text_target is not None:
|
|
if hasattr(self, "_switch_to_target_mode"):
|
|
self._switch_to_target_mode()
|
|
target_encodings = self._encode_plus(
|
|
text=text_target,
|
|
text_pair=text_pair_target,
|
|
**kwargs,
|
|
)
|
|
# Leave back tokenizer in input mode
|
|
if hasattr(self, "_switch_to_input_mode"):
|
|
self._switch_to_input_mode()
|
|
|
|
if text_target is None:
|
|
return encodings
|
|
elif text is None:
|
|
return target_encodings
|
|
else:
|
|
encodings["labels"] = target_encodings["input_ids"]
|
|
return encodings
|
|
|
|
@add_end_docstrings(UDOP_ENCODE_KWARGS_DOCSTRING)
|
|
def call_boxes(
|
|
self,
|
|
text: TextInput | PreTokenizedInput | list[TextInput] | list[PreTokenizedInput],
|
|
text_pair: PreTokenizedInput | list[PreTokenizedInput] | None = None,
|
|
boxes: list[list[int]] | list[list[list[int]]] | None = None,
|
|
word_labels: list[int] | list[list[int]] | None = None,
|
|
add_special_tokens: bool = True,
|
|
padding: bool | str | PaddingStrategy = False,
|
|
truncation: bool | str | TruncationStrategy = None,
|
|
max_length: int | None = None,
|
|
stride: int = 0,
|
|
pad_to_multiple_of: int | None = None,
|
|
padding_side: str | None = None,
|
|
return_tensors: str | TensorType | None = None,
|
|
return_token_type_ids: bool | None = None,
|
|
return_attention_mask: bool | None = None,
|
|
return_overflowing_tokens: bool = False,
|
|
return_special_tokens_mask: bool = False,
|
|
return_offsets_mapping: bool = False,
|
|
return_length: bool = False,
|
|
verbose: bool = True,
|
|
**kwargs,
|
|
) -> BatchEncoding:
|
|
"""
|
|
Main method to tokenize and prepare for the model one or several sequence(s) or one or several pair(s) of
|
|
sequences with word-level normalized bounding boxes and optional labels.
|
|
|
|
Args:
|
|
text (`str`, `list[str]`, `list[list[str]]`):
|
|
The sequence or batch of sequences to be encoded. Each sequence can be a string, a list of strings
|
|
(words of a single example or questions of a batch of examples) or a list of list of strings (batch of
|
|
words).
|
|
text_pair (`list[str]`, `list[list[str]]`):
|
|
The sequence or batch of sequences to be encoded. Each sequence should be a list of strings
|
|
(pretokenized string).
|
|
boxes (`list[list[int]]`, `list[list[list[int]]]`):
|
|
Word-level bounding boxes. Each bounding box should be normalized to be on a 0-1000 scale.
|
|
word_labels (`list[int]`, `list[list[int]]`, *optional*):
|
|
Word-level integer labels (for token classification tasks such as FUNSD, CORD).
|
|
"""
|
|
|
|
# Input type checking for clearer error
|
|
def _is_valid_text_input(t):
|
|
if isinstance(t, str):
|
|
# Strings are fine
|
|
return True
|
|
elif isinstance(t, (list, tuple)):
|
|
# List are fine as long as they are...
|
|
if len(t) == 0:
|
|
# ... empty
|
|
return True
|
|
elif isinstance(t[0], str):
|
|
# ... list of strings
|
|
return True
|
|
elif isinstance(t[0], (list, tuple)):
|
|
# ... list with an empty list or with a list of strings
|
|
return len(t[0]) == 0 or isinstance(t[0][0], str)
|
|
else:
|
|
return False
|
|
else:
|
|
return False
|
|
|
|
if text_pair is not None:
|
|
# in case text + text_pair are provided, text = questions, text_pair = words
|
|
if not _is_valid_text_input(text):
|
|
raise ValueError("text input must of type `str` (single example) or `list[str]` (batch of examples). ")
|
|
if not isinstance(text_pair, (list, tuple)):
|
|
raise ValueError(
|
|
"words must of type `list[str]` (single pretokenized example), "
|
|
"or `list[list[str]]` (batch of pretokenized examples)."
|
|
)
|
|
else:
|
|
# in case only text is provided => must be words
|
|
if not isinstance(text, (list, tuple)):
|
|
raise ValueError(
|
|
"Words must of type `list[str]` (single pretokenized example), "
|
|
"or `list[list[str]]` (batch of pretokenized examples)."
|
|
)
|
|
|
|
if text_pair is not None:
|
|
is_batched = isinstance(text, (list, tuple))
|
|
else:
|
|
is_batched = isinstance(text, (list, tuple)) and text and isinstance(text[0], (list, tuple))
|
|
|
|
words = text if text_pair is None else text_pair
|
|
if boxes is None:
|
|
raise ValueError("You must provide corresponding bounding boxes")
|
|
if is_batched:
|
|
if len(words) != len(boxes):
|
|
raise ValueError("You must provide words and boxes for an equal amount of examples")
|
|
for words_example, boxes_example in zip(words, boxes):
|
|
if len(words_example) != len(boxes_example):
|
|
raise ValueError("You must provide as many words as there are bounding boxes")
|
|
else:
|
|
if len(words) != len(boxes):
|
|
raise ValueError("You must provide as many words as there are bounding boxes")
|
|
|
|
if is_batched:
|
|
if text_pair is not None and len(text) != len(text_pair):
|
|
raise ValueError(
|
|
f"batch length of `text`: {len(text)} does not match batch length of `text_pair`:"
|
|
f" {len(text_pair)}."
|
|
)
|
|
batch_text_or_text_pairs = list(zip(text, text_pair)) if text_pair is not None else text
|
|
is_pair = bool(text_pair is not None)
|
|
return self.batch_encode_plus_boxes(
|
|
batch_text_or_text_pairs=batch_text_or_text_pairs,
|
|
is_pair=is_pair,
|
|
boxes=boxes,
|
|
word_labels=word_labels,
|
|
add_special_tokens=add_special_tokens,
|
|
padding=padding,
|
|
truncation=truncation,
|
|
max_length=max_length,
|
|
stride=stride,
|
|
pad_to_multiple_of=pad_to_multiple_of,
|
|
padding_side=padding_side,
|
|
return_tensors=return_tensors,
|
|
return_token_type_ids=return_token_type_ids,
|
|
return_attention_mask=return_attention_mask,
|
|
return_overflowing_tokens=return_overflowing_tokens,
|
|
return_special_tokens_mask=return_special_tokens_mask,
|
|
return_offsets_mapping=return_offsets_mapping,
|
|
return_length=return_length,
|
|
verbose=verbose,
|
|
**kwargs,
|
|
)
|
|
else:
|
|
return self.encode_plus_boxes(
|
|
text=text,
|
|
text_pair=text_pair,
|
|
boxes=boxes,
|
|
word_labels=word_labels,
|
|
add_special_tokens=add_special_tokens,
|
|
padding=padding,
|
|
truncation=truncation,
|
|
max_length=max_length,
|
|
stride=stride,
|
|
pad_to_multiple_of=pad_to_multiple_of,
|
|
padding_side=padding_side,
|
|
return_tensors=return_tensors,
|
|
return_token_type_ids=return_token_type_ids,
|
|
return_attention_mask=return_attention_mask,
|
|
return_overflowing_tokens=return_overflowing_tokens,
|
|
return_special_tokens_mask=return_special_tokens_mask,
|
|
return_offsets_mapping=return_offsets_mapping,
|
|
return_length=return_length,
|
|
verbose=verbose,
|
|
**kwargs,
|
|
)
|
|
|
|
def tokenize(self, text: str, pair: str | None = None, add_special_tokens: bool = False, **kwargs) -> list[str]:
|
|
batched_input = [(text, pair)] if pair else [text]
|
|
|
|
self._tokenizer.encode_special_tokens = kwargs.pop("split_special_tokens", self.split_special_tokens)
|
|
|
|
encodings = self._tokenizer.encode_batch(
|
|
batched_input, add_special_tokens=add_special_tokens, is_pretokenized=False, **kwargs
|
|
)
|
|
|
|
return encodings[0].tokens
|
|
|
|
def batch_encode_plus_boxes(
|
|
self,
|
|
batch_text_or_text_pairs: list[TextInput] | list[TextInputPair] | list[PreTokenizedInput],
|
|
is_pair: bool | None = None,
|
|
boxes: list[list[list[int]]] | None = None,
|
|
word_labels: list[list[int]] | None = None,
|
|
add_special_tokens: bool = True,
|
|
padding: bool | str | PaddingStrategy = False,
|
|
truncation: bool | str | TruncationStrategy = None,
|
|
max_length: int | None = None,
|
|
stride: int = 0,
|
|
is_split_into_words: bool = False,
|
|
pad_to_multiple_of: int | None = None,
|
|
padding_side: str | None = None,
|
|
return_tensors: str | TensorType | None = None,
|
|
return_token_type_ids: bool | None = None,
|
|
return_attention_mask: bool | None = None,
|
|
return_overflowing_tokens: bool = False,
|
|
return_special_tokens_mask: bool = False,
|
|
return_offsets_mapping: bool = False,
|
|
return_length: bool = False,
|
|
verbose: bool = True,
|
|
**kwargs,
|
|
) -> BatchEncoding:
|
|
"""
|
|
Tokenize and prepare for the model a list of sequences or a list of pairs of sequences.
|
|
|
|
<Tip warning={true}>
|
|
|
|
This method is deprecated, `__call__` should be used instead.
|
|
|
|
</Tip>
|
|
|
|
Args:
|
|
batch_text_or_text_pairs (`list[str]`, `list[tuple[str, str]]`, `list[list[str]]`, `list[tuple[list[str], list[str]]]`, and for not-fast tokenizers, also `list[list[int]]`, `list[tuple[list[int], list[int]]]`):
|
|
Batch of sequences or pair of sequences to be encoded. This can be a list of
|
|
string/string-sequences/int-sequences or a list of pair of string/string-sequences/int-sequence (see
|
|
details in `encode_plus`).
|
|
"""
|
|
|
|
# Backward compatibility for 'truncation_strategy', 'pad_to_max_length'
|
|
padding_strategy, truncation_strategy, max_length, kwargs = self._get_padding_truncation_strategies(
|
|
padding=padding,
|
|
truncation=truncation,
|
|
max_length=max_length,
|
|
pad_to_multiple_of=pad_to_multiple_of,
|
|
verbose=verbose,
|
|
**kwargs,
|
|
)
|
|
|
|
return self._batch_encode_plus_boxes(
|
|
batch_text_or_text_pairs=batch_text_or_text_pairs,
|
|
is_pair=is_pair,
|
|
boxes=boxes,
|
|
word_labels=word_labels,
|
|
add_special_tokens=add_special_tokens,
|
|
padding_strategy=padding_strategy,
|
|
truncation_strategy=truncation_strategy,
|
|
max_length=max_length,
|
|
stride=stride,
|
|
is_split_into_words=is_split_into_words,
|
|
pad_to_multiple_of=pad_to_multiple_of,
|
|
padding_side=padding_side,
|
|
return_tensors=return_tensors,
|
|
return_token_type_ids=return_token_type_ids,
|
|
return_attention_mask=return_attention_mask,
|
|
return_overflowing_tokens=return_overflowing_tokens,
|
|
return_special_tokens_mask=return_special_tokens_mask,
|
|
return_offsets_mapping=return_offsets_mapping,
|
|
return_length=return_length,
|
|
verbose=verbose,
|
|
**kwargs,
|
|
)
|
|
|
|
def _batch_encode_plus_boxes(
|
|
self,
|
|
batch_text_or_text_pairs: list[TextInput] | list[TextInputPair] | list[PreTokenizedInput],
|
|
is_pair: bool | None = None,
|
|
boxes: list[list[list[int]]] | None = None,
|
|
word_labels: list[list[int]] | None = None,
|
|
add_special_tokens: bool = True,
|
|
padding_strategy: PaddingStrategy = PaddingStrategy.DO_NOT_PAD,
|
|
truncation_strategy: TruncationStrategy = TruncationStrategy.DO_NOT_TRUNCATE,
|
|
max_length: int | None = None,
|
|
stride: int = 0,
|
|
pad_to_multiple_of: int | None = None,
|
|
padding_side: str | None = None,
|
|
return_tensors: str | None = None,
|
|
return_token_type_ids: bool | None = None,
|
|
return_attention_mask: bool | None = None,
|
|
return_overflowing_tokens: bool = False,
|
|
return_special_tokens_mask: bool = False,
|
|
return_offsets_mapping: bool = False,
|
|
return_length: bool = False,
|
|
verbose: bool = True,
|
|
**kwargs,
|
|
) -> BatchEncoding:
|
|
if not isinstance(batch_text_or_text_pairs, list):
|
|
raise TypeError(f"batch_text_or_text_pairs has to be a list (got {type(batch_text_or_text_pairs)})")
|
|
|
|
# Set the truncation and padding strategy and restore the initial configuration
|
|
self.set_truncation_and_padding(
|
|
padding_strategy=padding_strategy,
|
|
truncation_strategy=truncation_strategy,
|
|
max_length=max_length,
|
|
stride=stride,
|
|
pad_to_multiple_of=pad_to_multiple_of,
|
|
padding_side=padding_side,
|
|
)
|
|
|
|
if is_pair:
|
|
batch_text_or_text_pairs = [(text.split(), text_pair) for text, text_pair in batch_text_or_text_pairs]
|
|
|
|
encodings = self._tokenizer.encode_batch(
|
|
batch_text_or_text_pairs,
|
|
add_special_tokens=add_special_tokens,
|
|
is_pretokenized=True, # we set this to True as LayoutLMv2 always expects pretokenized inputs
|
|
)
|
|
|
|
# Convert encoding to dict
|
|
# `Tokens` has type: tuple[
|
|
# list[dict[str, list[list[int]]]] or list[dict[str, 2D-Tensor]],
|
|
# list[EncodingFast]
|
|
# ]
|
|
# with nested dimensions corresponding to batch, overflows, sequence length
|
|
tokens_and_encodings = [
|
|
self._convert_encoding(
|
|
encoding=encoding,
|
|
return_token_type_ids=return_token_type_ids,
|
|
return_attention_mask=return_attention_mask,
|
|
return_overflowing_tokens=return_overflowing_tokens,
|
|
return_special_tokens_mask=return_special_tokens_mask,
|
|
return_offsets_mapping=True
|
|
if word_labels is not None
|
|
else return_offsets_mapping, # we use offsets to create the labels
|
|
return_length=return_length,
|
|
verbose=verbose,
|
|
)
|
|
for encoding in encodings
|
|
]
|
|
|
|
# Convert the output to have dict[list] from list[dict] and remove the additional overflows dimension
|
|
# From (variable) shape (batch, overflows, sequence length) to ~ (batch * overflows, sequence length)
|
|
# (we say ~ because the number of overflow varies with the example in the batch)
|
|
#
|
|
# To match each overflowing sample with the original sample in the batch
|
|
# we add an overflow_to_sample_mapping array (see below)
|
|
sanitized_tokens = {}
|
|
for key in tokens_and_encodings[0][0]:
|
|
stack = [e for item, _ in tokens_and_encodings for e in item[key]]
|
|
sanitized_tokens[key] = stack
|
|
sanitized_encodings = [e for _, item in tokens_and_encodings for e in item]
|
|
|
|
# If returning overflowing tokens, we need to return a mapping
|
|
# from the batch idx to the original sample
|
|
if return_overflowing_tokens:
|
|
overflow_to_sample_mapping = []
|
|
for i, (toks, _) in enumerate(tokens_and_encodings):
|
|
overflow_to_sample_mapping += [i] * len(toks["input_ids"])
|
|
sanitized_tokens["overflow_to_sample_mapping"] = overflow_to_sample_mapping
|
|
|
|
for input_ids in sanitized_tokens["input_ids"]:
|
|
self._eventual_warn_about_too_long_sequence(input_ids, max_length, verbose)
|
|
|
|
# create the token boxes
|
|
token_boxes = []
|
|
for batch_index in range(len(sanitized_tokens["input_ids"])):
|
|
if return_overflowing_tokens:
|
|
original_index = sanitized_tokens["overflow_to_sample_mapping"][batch_index]
|
|
else:
|
|
original_index = batch_index
|
|
token_boxes_example = []
|
|
for id, sequence_id, word_id in zip(
|
|
sanitized_tokens["input_ids"][batch_index],
|
|
sanitized_encodings[batch_index].sequence_ids,
|
|
sanitized_encodings[batch_index].word_ids,
|
|
):
|
|
if word_id is not None:
|
|
if is_pair and sequence_id == 0:
|
|
token_boxes_example.append(self.pad_token_box)
|
|
else:
|
|
token_boxes_example.append(boxes[original_index][word_id])
|
|
else:
|
|
if id == self.sep_token_id:
|
|
token_boxes_example.append(self.sep_token_box)
|
|
elif id == self.pad_token_id:
|
|
token_boxes_example.append(self.pad_token_box)
|
|
else:
|
|
raise ValueError("Id not recognized")
|
|
token_boxes.append(token_boxes_example)
|
|
|
|
sanitized_tokens["bbox"] = token_boxes
|
|
|
|
# optionally, create the labels
|
|
if word_labels is not None:
|
|
labels = []
|
|
for batch_index in range(len(sanitized_tokens["input_ids"])):
|
|
if return_overflowing_tokens:
|
|
original_index = sanitized_tokens["overflow_to_sample_mapping"][batch_index]
|
|
else:
|
|
original_index = batch_index
|
|
labels_example = []
|
|
previous_token_empty = False
|
|
for id, offset, word_id in zip(
|
|
sanitized_tokens["input_ids"][batch_index],
|
|
sanitized_tokens["offset_mapping"][batch_index],
|
|
sanitized_encodings[batch_index].word_ids,
|
|
):
|
|
if word_id is not None:
|
|
if self.only_label_first_subword:
|
|
if offset[0] == 0 and not previous_token_empty:
|
|
# Use the real label id for the first token of the word, and padding ids for the remaining tokens
|
|
labels_example.append(word_labels[original_index][word_id])
|
|
else:
|
|
labels_example.append(self.pad_token_label)
|
|
else:
|
|
labels_example.append(word_labels[original_index][word_id])
|
|
if self.decode(id) == "":
|
|
previous_token_empty = True
|
|
else:
|
|
previous_token_empty = False
|
|
else:
|
|
labels_example.append(self.pad_token_label)
|
|
labels.append(labels_example)
|
|
|
|
sanitized_tokens["labels"] = labels
|
|
# finally, remove offsets if the user didn't want them
|
|
if not return_offsets_mapping:
|
|
del sanitized_tokens["offset_mapping"]
|
|
|
|
return BatchEncoding(sanitized_tokens, sanitized_encodings, tensor_type=return_tensors)
|
|
|
|
def _encode_plus_boxes(
|
|
self,
|
|
text: TextInput | PreTokenizedInput,
|
|
text_pair: PreTokenizedInput | None = None,
|
|
boxes: list[list[int]] | None = None,
|
|
word_labels: list[int] | None = None,
|
|
add_special_tokens: bool = True,
|
|
padding_strategy: PaddingStrategy = PaddingStrategy.DO_NOT_PAD,
|
|
truncation_strategy: TruncationStrategy = TruncationStrategy.DO_NOT_TRUNCATE,
|
|
max_length: int | None = None,
|
|
stride: int = 0,
|
|
pad_to_multiple_of: int | None = None,
|
|
padding_side: str | None = None,
|
|
return_tensors: bool | None = None,
|
|
return_token_type_ids: bool | None = None,
|
|
return_attention_mask: bool | None = None,
|
|
return_overflowing_tokens: bool = False,
|
|
return_special_tokens_mask: bool = False,
|
|
return_offsets_mapping: bool = False,
|
|
return_length: bool = False,
|
|
verbose: bool = True,
|
|
**kwargs,
|
|
) -> BatchEncoding:
|
|
# make it a batched input
|
|
# 2 options:
|
|
# 1) only text, in case text must be a list of str
|
|
# 2) text + text_pair, in which case text = str and text_pair a list of str
|
|
batched_input = [(text, text_pair)] if text_pair else [text]
|
|
batched_boxes = [boxes]
|
|
batched_word_labels = [word_labels] if word_labels is not None else None
|
|
batched_output = self._batch_encode_plus_boxes(
|
|
batched_input,
|
|
is_pair=bool(text_pair is not None),
|
|
boxes=batched_boxes,
|
|
word_labels=batched_word_labels,
|
|
add_special_tokens=add_special_tokens,
|
|
padding_strategy=padding_strategy,
|
|
truncation_strategy=truncation_strategy,
|
|
max_length=max_length,
|
|
stride=stride,
|
|
pad_to_multiple_of=pad_to_multiple_of,
|
|
padding_side=padding_side,
|
|
return_tensors=return_tensors,
|
|
return_token_type_ids=return_token_type_ids,
|
|
return_attention_mask=return_attention_mask,
|
|
return_overflowing_tokens=return_overflowing_tokens,
|
|
return_special_tokens_mask=return_special_tokens_mask,
|
|
return_offsets_mapping=return_offsets_mapping,
|
|
return_length=return_length,
|
|
verbose=verbose,
|
|
**kwargs,
|
|
)
|
|
|
|
# Return tensor is None, then we can remove the leading batch axis
|
|
# Overflowing tokens are returned as a batch of output so we keep them in this case
|
|
if return_tensors is None and not return_overflowing_tokens:
|
|
batched_output = BatchEncoding(
|
|
{
|
|
key: value[0] if len(value) > 0 and isinstance(value[0], list) else value
|
|
for key, value in batched_output.items()
|
|
},
|
|
batched_output.encodings,
|
|
)
|
|
|
|
self._eventual_warn_about_too_long_sequence(batched_output["input_ids"], max_length, verbose)
|
|
|
|
return batched_output
|
|
|
|
def encode_boxes(
|
|
self,
|
|
text: TextInput | PreTokenizedInput | EncodedInput,
|
|
text_pair: TextInput | PreTokenizedInput | EncodedInput | None = None,
|
|
boxes: list[list[int]] | None = None,
|
|
word_labels: list[list[int]] | None = None,
|
|
add_special_tokens: bool = True,
|
|
padding: bool | str | PaddingStrategy = False,
|
|
truncation: bool | str | TruncationStrategy = None,
|
|
max_length: int | None = None,
|
|
stride: int = 0,
|
|
return_tensors: str | TensorType | None = None,
|
|
**kwargs,
|
|
) -> list[int]:
|
|
"""
|
|
Args:
|
|
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))`.
|
|
text (`str`, `list[str]` or `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 (`str`, `list[str]` or `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_boxes(
|
|
text,
|
|
text_pair=text_pair,
|
|
boxes=boxes,
|
|
word_labels=word_labels,
|
|
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 encode_plus_boxes(
|
|
self,
|
|
text: TextInput | PreTokenizedInput,
|
|
text_pair: PreTokenizedInput | None = None,
|
|
boxes: list[list[int]] | None = None,
|
|
word_labels: list[list[int]] | None = None,
|
|
add_special_tokens: bool = True,
|
|
padding: bool | str | PaddingStrategy = False,
|
|
truncation: bool | str | TruncationStrategy = None,
|
|
max_length: int | None = None,
|
|
stride: int = 0,
|
|
is_split_into_words: bool = False,
|
|
pad_to_multiple_of: int | None = None,
|
|
padding_side: str | None = None,
|
|
return_tensors: str | TensorType | None = None,
|
|
return_token_type_ids: bool | None = None,
|
|
return_attention_mask: bool | None = None,
|
|
return_overflowing_tokens: bool = False,
|
|
return_special_tokens_mask: bool = False,
|
|
return_offsets_mapping: bool = False,
|
|
return_length: bool = False,
|
|
verbose: bool = True,
|
|
**kwargs,
|
|
) -> BatchEncoding:
|
|
"""
|
|
Tokenize and prepare for the model a sequence or a pair of sequences.
|
|
|
|
<Tip warning={true}>
|
|
|
|
This method is deprecated, `__call__` should be used instead.
|
|
|
|
</Tip>
|
|
|
|
Args:
|
|
text (`str`, `list[str]` or (for non-fast tokenizers) `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 (`str`, `list[str]` or `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).
|
|
"""
|
|
|
|
# Backward compatibility for 'truncation_strategy', 'pad_to_max_length'
|
|
padding_strategy, truncation_strategy, max_length, kwargs = self._get_padding_truncation_strategies(
|
|
padding=padding,
|
|
truncation=truncation,
|
|
max_length=max_length,
|
|
pad_to_multiple_of=pad_to_multiple_of,
|
|
verbose=verbose,
|
|
**kwargs,
|
|
)
|
|
|
|
return self._encode_plus_boxes(
|
|
text=text,
|
|
text_pair=text_pair,
|
|
boxes=boxes,
|
|
word_labels=word_labels,
|
|
add_special_tokens=add_special_tokens,
|
|
padding_strategy=padding_strategy,
|
|
truncation_strategy=truncation_strategy,
|
|
max_length=max_length,
|
|
stride=stride,
|
|
is_split_into_words=is_split_into_words,
|
|
pad_to_multiple_of=pad_to_multiple_of,
|
|
padding_side=padding_side,
|
|
return_tensors=return_tensors,
|
|
return_token_type_ids=return_token_type_ids,
|
|
return_attention_mask=return_attention_mask,
|
|
return_overflowing_tokens=return_overflowing_tokens,
|
|
return_special_tokens_mask=return_special_tokens_mask,
|
|
return_offsets_mapping=return_offsets_mapping,
|
|
return_length=return_length,
|
|
verbose=verbose,
|
|
**kwargs,
|
|
)
|
|
|
|
def _pad(
|
|
self,
|
|
encoded_inputs: dict[str, EncodedInput] | BatchEncoding,
|
|
max_length: int | None = None,
|
|
padding_strategy: PaddingStrategy = PaddingStrategy.DO_NOT_PAD,
|
|
pad_to_multiple_of: int | None = None,
|
|
padding_side: str | None = None,
|
|
return_attention_mask: bool | None = None,
|
|
) -> dict:
|
|
"""
|
|
Pad encoded inputs (on left/right and up to predefined length or max length in the batch)
|
|
|
|
Args:
|
|
encoded_inputs:
|
|
Dictionary of tokenized inputs (`list[int]`) or batch of tokenized inputs (`list[list[int]]`).
|
|
max_length: maximum length of the returned list and optionally padding length (see below).
|
|
Will truncate by taking into account the special tokens.
|
|
padding_strategy: PaddingStrategy to use for padding.
|
|
|
|
- PaddingStrategy.LONGEST Pad to the longest sequence in the batch
|
|
- PaddingStrategy.MAX_LENGTH: Pad to the max length (default)
|
|
- PaddingStrategy.DO_NOT_PAD: Do not pad
|
|
The tokenizer padding sides are defined in self.padding_side:
|
|
|
|
- 'left': pads on the left of the sequences
|
|
- 'right': pads on the right of the sequences
|
|
pad_to_multiple_of: (optional) Integer if set will pad the sequence to a multiple of the provided value.
|
|
This is especially useful to enable the use of Tensor Core on NVIDIA hardware with compute capability
|
|
`>= 7.5` (Volta).
|
|
padding_side (`str`, *optional*):
|
|
The side on which the model should have padding applied. Should be selected between ['right', 'left'].
|
|
Default value is picked from the class attribute of the same name.
|
|
return_attention_mask:
|
|
(optional) Set to False to avoid returning attention mask (default: set to model specifics)
|
|
"""
|
|
# Load from model defaults
|
|
if return_attention_mask is None:
|
|
return_attention_mask = "attention_mask" in self.model_input_names
|
|
|
|
required_input = encoded_inputs[self.model_input_names[0]]
|
|
|
|
if padding_strategy == PaddingStrategy.LONGEST:
|
|
max_length = len(required_input)
|
|
|
|
if max_length is not None and pad_to_multiple_of is not None and (max_length % pad_to_multiple_of != 0):
|
|
max_length = ((max_length // pad_to_multiple_of) + 1) * pad_to_multiple_of
|
|
|
|
needs_to_be_padded = padding_strategy != PaddingStrategy.DO_NOT_PAD and len(required_input) != max_length
|
|
|
|
# Initialize attention mask if not present.
|
|
if return_attention_mask and "attention_mask" not in encoded_inputs:
|
|
encoded_inputs["attention_mask"] = [1] * len(required_input)
|
|
|
|
if needs_to_be_padded:
|
|
difference = max_length - len(required_input)
|
|
padding_side = padding_side if padding_side is not None else self.padding_side
|
|
if padding_side == "right":
|
|
if return_attention_mask:
|
|
encoded_inputs["attention_mask"] = encoded_inputs["attention_mask"] + [0] * difference
|
|
if "token_type_ids" in encoded_inputs:
|
|
encoded_inputs["token_type_ids"] = (
|
|
encoded_inputs["token_type_ids"] + [self.pad_token_type_id] * difference
|
|
)
|
|
if "bbox" in encoded_inputs:
|
|
encoded_inputs["bbox"] = encoded_inputs["bbox"] + [self.pad_token_box] * difference
|
|
if "labels" in encoded_inputs:
|
|
encoded_inputs["labels"] = encoded_inputs["labels"] + [self.pad_token_label] * difference
|
|
if "special_tokens_mask" in encoded_inputs:
|
|
encoded_inputs["special_tokens_mask"] = encoded_inputs["special_tokens_mask"] + [1] * difference
|
|
encoded_inputs[self.model_input_names[0]] = required_input + [self.pad_token_id] * difference
|
|
elif padding_side == "left":
|
|
if return_attention_mask:
|
|
encoded_inputs["attention_mask"] = [0] * difference + encoded_inputs["attention_mask"]
|
|
if "token_type_ids" in encoded_inputs:
|
|
encoded_inputs["token_type_ids"] = [self.pad_token_type_id] * difference + encoded_inputs[
|
|
"token_type_ids"
|
|
]
|
|
if "bbox" in encoded_inputs:
|
|
encoded_inputs["bbox"] = [self.pad_token_box] * difference + encoded_inputs["bbox"]
|
|
if "labels" in encoded_inputs:
|
|
encoded_inputs["labels"] = [self.pad_token_label] * difference + encoded_inputs["labels"]
|
|
if "special_tokens_mask" in encoded_inputs:
|
|
encoded_inputs["special_tokens_mask"] = [1] * difference + encoded_inputs["special_tokens_mask"]
|
|
encoded_inputs[self.model_input_names[0]] = [self.pad_token_id] * difference + required_input
|
|
else:
|
|
raise ValueError("Invalid padding strategy:" + str(padding_side))
|
|
|
|
return encoded_inputs
|
|
|
|
def build_inputs_with_special_tokens(
|
|
self, token_ids_0: list[int], token_ids_1: list[int] | None = None
|
|
) -> list[int]:
|
|
"""
|
|
Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
|
|
adding special tokens. An XLM-RoBERTa sequence has the following format:
|
|
|
|
- single sequence: `<s> X </s>`
|
|
- pair of sequences: `<s> A </s></s> B </s>`
|
|
|
|
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.
|
|
"""
|
|
|
|
if token_ids_1 is None:
|
|
return token_ids_0 + [self.sep_token_id]
|
|
sep = [self.sep_token_id]
|
|
return token_ids_0 + sep + token_ids_1 + sep
|
|
|
|
def create_token_type_ids_from_sequences(
|
|
self, token_ids_0: list[int], token_ids_1: list[int] | None = None
|
|
) -> list[int]:
|
|
"""
|
|
Create a mask from the two sequences passed to be used in a sequence-pair classification task. XLM-RoBERTa 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.
|
|
|
|
"""
|
|
|
|
sep = [self.sep_token_id]
|
|
|
|
if token_ids_1 is None:
|
|
return len(token_ids_0 + sep) * [0]
|
|
return len(token_ids_0 + sep + token_ids_1 + sep) * [0]
|
|
|
|
def save_vocabulary(self, save_directory: str, filename_prefix: str | None = None) -> tuple[str]:
|
|
"""
|
|
Save the tokenizer vocabulary files. For TokenizersBackend, the tokenizer.json file is saved
|
|
by the base class. This method returns an empty tuple since we only use tokenizer.json.
|
|
"""
|
|
# The base class handles saving tokenizer.json in _save_pretrained
|
|
# We don't need to save vocab_file since we only use tokenizer.json
|
|
return ()
|
|
|
|
|
|
__all__ = ["UdopTokenizer"]
|